Primary pediatric musculoskeletal malignancies are relatively uncommon, representing 15% of pediatric malignancies.18 Osteosarcoma and Ewing's sarcoma account for 85% of the annual cases of pediatric musculoskeletal malignancies per year. Though these tumors can occur in the axial skeleton, most originate in the extremities. Accurate staging of the extent of disease is extremely important in the development of a sarcoma treatment plan to minimize treatment morbidity and maximize survival.17 Determinants of stage include size, nodal involvement, distant metastasis, and histologic grade. Histologic examination of biopsy material determines the tumor grade, while assessment for local extent, regional spread, and distant metastases is accomplished by a variety of radiographic imaging techniques. Local imaging examinations include computed tomography (CT) and magnetic resonance imaging (MRI).17 Pulmonary metastases are best screened by chest CT. Whole body technetium bone scans are routinely performed to evaluate patients for secondary bone lesions in bone sarcomas and are especially helpful in osteosarcoma.10,19
Positron emission tomography (PET) is an imaging technique that reflects tissue metabolic activity rather than anatomic detail. Malignant tumors have an increased rate of glycolysis compared with benign tissue and show increased uptake of fluorine-18-fluoro-2-deoxy D-glucose (FDG) on PET. First used to study malignant gliomas, PET has subsequently been used to evaluate many other malignancies.2-7,19 The use of PET with FDG to assess human extremity musculoskeletal tumors was introduced in 1988.13 Subsequently, PET with FDG has been extensively investigated in the evaluation of musculoskeletal neoplasms, including diagnostic evaluation and staging, chemotherapeutic response, and surveillance.1,4,7,9,11,13-16 Orthopaedic oncologists are increasingly using PET technology in the initial work-up and staging of sarcomas and for monitoring therapy response.1,4,6,7,9
When we initially began using PET with FDG (beginning December 1994), the role of PET with FDG in sarcoma patients was unproven. We focused our use of PET routinely in the staging of patients diagnosed with high grade bone and soft tissue sarcomas. We directed specific attention to the presence or absence of occult, nonpulmo- nary metastases. We intended to ascertain whether PET could contribute to the treatment decisions, particularly the early identification of patients with occult Stage IV disease.
We asked whether PET with FDG as an initial staging tool could identify occult nonpulmonary metastases in pediatric patients newly diagnosed with either Ewing's or osteosarcoma. We then asked whether the detection of such metastases altered treatment decisions.
MATERIALS AND METHODS
We reviewed 55 patients referred to the senior author (JK) between December 1994 and November 2004. Based on history, physical examination, and plain radiographs, all patients (225) with lesions strongly suspected to be primary sarcomas of an extremity underwent CT and/or MRI of the primary site, chest radiographs, and chest CT as standard protocol. All patients with bone lesions underwent technetium bone scans and whole body PET with FDG. The inclusion criteria were defined by the Pediatric Oncology Group for clinical trials for pediatric bone sarcomas.7 These criteria specify age < 30 years for clinical trials involving osteosarcoma, Ewing's sarcoma or primitive neuroectodermal tumor. We included 55 patients under 29 years of age (range 6-29) diagnosed with Ewing's sarcoma/primitive neuroectodermal tumor (PNET) (17 patients), or osteosarcoma (38 patients). Gender distribution was nearly equal (28 male and 27 female patients). All patients underwent random iliac crest bone marrow sampling. All data were collected prospectively and retrospectively reviewed, and the study was approved by the institutional review board.
Prior to PET scanning, all patients were interviewed for relevant clinical history to supplement the referring physician office notes. Clinical history included dates of initial diagnosis, chemotherapy, radiation therapy, surgeries or recent infections, use of white or red blood cell stimulating drugs, diabetes screen, and contraindications for diazepam.
Patients were required to fast for more than 4 hours and instructed to hydrate with 24 to 30 ounces of water over a 3-hour interval before the scan. Blood glucose values were obtained upon patient's arrival to document euglycemia. All patients included in this study had blood glucose value below 120 mg. Patients were placed on a comfortable hospital bed and given a weight adjusted dose of crushed oral diazepam to reduce normal skeletal muscle activity. A dose of 12 to 20 mCi of FDG was injected intravenously after waiting 20 to 30 minutes for the oral diazepam to work. Furosemide was administered intravenously 45 minutes after FDG. The patient's intravenous adapter was removed and they were escorted within 10 minutes to the bathroom for an initial emptying of the bladder. The patients were returned to the holding area where they waited 15 more minutes for the bladder to refill. Patients then voided one more time before scanning. Low grade bladder activity proves useful as an anatomic marker for pelvic pathology and eliminates the need for urinary bladder catheterization. Approximately 80% of our patients can tolerate 48 minutes of whole body imaging with this diuresis protocol. This protocol is designed to reduce the overall bladder radiation dose and is currently being investigated at our institution. The remaining patient studies are acquired in two data sets with the first covering the chest-abdomen and the second covering the abdomen-pelvis with one bed overlap.
All PET scans were obtained on a Siemens/CTI ECAT 951-R whole body PET scanner (Siemens Medical Systems, Hoffman Estates, IL; CTI, Knoxville, TN). A total of 31 contiguous image planes were generated in each field of view, each with 3.375-mm plane spacing. Images were reconstructed by sending sinograms to the array processor and back projector, which performed parallel beam transaxial reconstruction. Reconstruction parameters included a Hann spatial filter with a 0.4 cutoff frequency (cycles/pixel) and a 128 × 128 pixels image size. Images were x-y-z transaxial (z) direction. Average axial resolution with the study time frame was 4.7-mm FWHM and measured sensitivity average 110,000 counts/microCurie/milliliter.
One radiologist (JZ) reviewed the images immediately after the completion of each study for determination of technical image quality. Qualitative assessment of the images was accomplished using a Sun SPARC station 10 computer (Sun Microsystems, Mountain View, CA) with high-resolution 16-inch monitors. Images were manipulated and displayed using the ECAT Version 6.5B system software (Siemens Medical Systems; CTI). Images were then processed into transverse, coronal, and sagittal and normalized whole body projection and volume views.
Patients with primary tumor sites in the upper extremity, chest, abdomen, or pelvis were scanned with the arms folded across their abdomen. Imaging was started at the neck working towards the pelvis for eight beds, acquired 6 minutes each, for a total of 48 minutes. Lower extremity primary sites required a slightly different protocol. In these patients, the lower extremities were imaged first after the FDG uptake period. Patients were wheeled into the scan room and positioned supine and feet first into the gantry. Imaging was set up for 10 beds, acquired 4 minutes each, for 40 total minutes. The technologist then administered the furosemide diuresis protocol and the patient would void twice over a period of 20 minutes. This eliminated muscle activation artifacts of the lower extremities caused by walking to the bathroom if the torso scan was attempted first. Chest, abdomen, and pelvis scans were then completed with the same parameters as described previously with upper extremity primary sites.
Image interpretation at the time of this study was based on visual analysis. Standard uptake values were not routinely measured or recorded during 1994 to 2000. The tumor activity was compared with muscle, mediastinal, and background activity. Activity equal to or less than muscle activity was considered benign. Activity greater than muscle was considered malignant, with more intense activity correlating to high grade tumors and less intense activity to lower grade lesions. Heterogeneous up- take was considered consistent with tumor necrosis, dedifferentiation, or a mixed tumor with varying cellularity based on cyst formation or fibrosis.
Physiologic variation in FDG activity was seen in muscle, gastrointestinal, urinary, and cardiac systems. Transmission scans for attenuation correction were only used in cases of intense myocardial uptake to eliminate streak artifacts that limit accurate evaluation of adjacent lung and liver fields. Delayed scans performed at longer imaging times of 10 minutes per bed over designated areas were occasionally used to differentiate benign from malignant disease.
All patients had previous CT scans of the chest and MRI and/or CT scans of the tumor bed for correlation to the PET findings, using GE 1.5 Tesla magnets and GE helical scanners (General Electric Medical Systems, Milwaukee, WI, USA). These scans had already been interpreted without the results of the PET scan. Any clinically suspicious or PET positive areas were additionally investigated in an attempt to accurately stage these patients and direct surgical, chemotherapeutic, and radiation oncology decisions. In several cases, a positive PET scan indicated a second MRI and CT scan to additionally evaluate the positive finding. Usually the positive PET finding would also be identified with these tailored high resolution studies. In such cases, the suspected metastatic lesion would be biopsied to document distant disease. If the followup MRI and CT scan remained negative, a determination was made as to the likelihood of a false positive PET scan. In this instance, the positive PET finding would usually be followed clinically or with a followup CT/MRI scan. Patient survival data was determined from the Institutional Tumor Registry. Our Institutional Tumor Registry consists of a dedicated staff of cancer data registrars who collate and abstract all cancer cases (diagnosis, stage, treatment, and followup) for annual submission to the North Carolina Cancer Database. Annual followup data is accrued from treating physicians and survival/mortality is submitted as well.
Among the 55 patients all primary tumors showed increased uptake of FDG on PET scan. PET detected metastases in 12 of 55 (22%) of these patients at initial staging. Eight of the 12 patients with metastases (67%) had nonpulmonary metastases. Seven of the eight patients with nonpulmonary metastasis had bony metastases, of which five had a negative bone scan (four Ewing's/PNET and one osteosarcoma). In all five cases with a negative bone scan (for metastasis), the primary lesion was visible on the PET scan.
Based upon added information from PET at initial staging, four of the 55 (7%) patients in this series were up- staged to Stage IV based upon the detection of occult nonpulmonary metastases. These four patients had their treatment regimens modified based upon this new information. Specifically, for the Ewing's/PNET patients local irradiation in lieu of surgical resection was offered to musculoskeletal sites. For the single osteosarcoma patient, local resection was offered to both musculoskeletal sites (distal femur, sacroiliac). The additional four patients were identified as having both pulmonary and nonpulmonary metastases, and the information from PET did not modify their initial stage at diagnosis nor their treatment recommendations. Three of 17 (18%) Ewing's/PNET and one of 38 (3%) osteosarcomas were upstaged because of the PET results.
The cost of PET scanning all patients included in this study was approximately $110,000 ($2000.00 × 55). There was one false negative PET scan with a positive bone scan in the case of an osteogenic osteosarcoma. All random iliac crest bone marrows in patients with occult nonpulmonary metastases were negative. All patients who were diagnosed with Stage IV Ewing's sarcoma and osteosarcoma have expired.
PET with FDG is a relatively new diagnostic modality in which malignant tumors have an increased rate of glycolysis. Given the potential to detect otherwise occult metastasis, we asked whether PET with FDG as an initial staging tool could identify occult nonpulmonary metastases in pediatric patients newly diagnosed with either Ewing's or osteosarcoma. We then asked whether the detection of such metastases altered treatment decisions.
We note several limitations. This study is limited by its small size and is underpowered to account for confounding variables such as tumor type. Data accumulation was initiated during the early phase of the senior author's practice and accrual was slow. The PET technology utilized during the study and subjective interpretive methods are considered to be below current state of the art, and have been supplanted at the authors' institution by high resolution PET/CT. Statistical manipulation of data is limited by lack of correlative axial scans on the negative PET scans. Acknowledging these limitations, a substantial number of Ewing's sarcoma patients had occult nonpulmonary metastases detected and are the focal point of the following discussion.
We found PET can detect asymptomatic (occult) metastases in sites not usually studied for sarcomas in general, and in particular can detect bone scan negative skeletal metastases in Ewing's sarcoma in 18% of the patients studied. The yield of PET versus technetium bone scan in identifying bone metastases was superior in only one osteosarcoma patient but four of the Ewing's/PNET patients. Given this information, we conclude PET is a useful adjunct in the initial staging of Ewing's sarcoma, but relatively noncontributory in patients diagnosed with osteosarcoma. One published report has suggested PET is most advantageous for purely lytic lesions, such as seen with Ewing's sarcoma, osteolytic osteosarcoma, and multiple myeloma, which may all have negative bone scans.3,12
The utility of new diagnostic procedures is based upon their impact on patient management, therapeutic strategies, and contribution of information provided by the new diagnostic procedure not otherwise provided by conventional studies. PET with FDG had historically seen limited use in the United States for orthopaedic applications given its relative expense, unavailability, and a lack of reimbursement by third party payers, but has undergone a substantial change in recent years.8 In this current study, only 4/55 (7%) of patients have had their treatment recommendation altered based upon information obtainable from PET alone. Three of 17 (18%) of Ewing's sarcoma patients avoided surgery. The relative cost savings of irradiation versus surgery were negligible.8 The other four patients with occult nonpulmonary metastases were already Stage IV on the basis of chest CT detection of pulmonary metastases as well.
The value of PET scanning lies in the visualization of high metabolic activity, particularly the metabolic activity of biologically active tumors such as sarcoma. PET scanning as performed during the study period does not have the same anatomic detail of CT or MRI scanning. Our current protocol is to obtain a combined PET and CT, which provides improved localization of abnormal FDG activity by fusing metabolic (PET) and anatomic (CT) data at the same setting. This may replace the need for a directed chest CT to stage these tumors; however, we cannot recommend substitution of PET for chest CT for detection of pulmonary metastases based on our current research.
With improved image resolution, ease of operation, and speed of whole body PET scanners and standardization of expected diagnostic imaging and treatment protocols, the utilization of PET is increasing at a rapid pace. While the advantages of the combined newest generation of PET/CT (or even PET/MRI) technologies may prove to be the future gold standard for imaging and tumor staging, they were not available or investigated in the present study.
For staging evaluations of pediatric bone sarcomas, the current study supports the use of PET scanning as an adjunct to the initial staging of patients afflicted with Ewing's sarcoma. The primary benefit to patients is the avoidance of surgery for control of local disease, as all Stage IV Ewing's sarcoma patients eventually died of disseminated disease. The utility of PET scanning to detect occult non- pulmonary metastases in osteosarcoma is minimal.
1. Conrad EU, Morgan HD, Vernon C, Schuetze SM, Eary JF. Fluorodeoxyglucose positron emission tomography scanning: basic principles and imaging of adult soft-tissue sarcomas. J Bone Joint Surg Am
. 2004:86(Suppl 2):98-104.
2. Di Chiro G, DeLaPaz RL, Brooks RA, Sokoloff L, Kornblith PL, Smith BH, Patronas NJ, Kufta CV, Kessler RM, Johnston GS, Manning RG, Wolf AP. Glucose utilization of cerebral gliomas measured by [18F] fluorodeoxyglucose and PET. Neurology
. 1982;32: 1323-1329.
3. Feldman F, van Heertum R, Manos C. 18FDG PET scanning of benign and malignant musculoskeletal lesion. Skeletal Radiol
. 2003; 32:201-208.
4. Folpe AL, Lyles RH, Sprouse JT, Conrad EU, Eary JF. (F-18) Fluorodeoxyglucose positron emission tomography as a predictor of pathologic grade and other prognostic variables in bone and soft tissue sarcoma. Clin Cancer Resh
5. Glaspy JA, Hawkins R, Hoh CK, Phelps ME. Use of positron emission tomography in oncology. Oncology
6. Gupta NC, Frick MP. Clinical applications of positron emission tomography in cancer. CA Cancer J Clin
7. Hawkins DS, Rajendran JG, Conrad EU, Bruckner JD, Eary JF. Evaluation of chemotherapy response in pediatric bone sarcomas by (F-18) fluorodeoxyglucose positron emission tomography. Cancer
8. Healthcare Financial Administration: Decision Memo for Medicare Coverage of FDG Positron Emission Tomography. Available at: http://www.cms.hhs.gov/mcd/viewdecisionmemo.asp
9. Ioannidis JPA, Lau J. 18F-FDG PET for the diagnosis and grading of soft tissue sarcoma: a meta-analysis. J Nucl Med
. 2003;44: 717-724.
10. Jager PL, Hoekstra HJ, Leeuw JA, van Der Graaf WT, de Vries EG, Piers D. Routine bone scintigraphy in primary staging of soft tissue sarcoma. Cancer
11. Jones DN, McCowage GB, Sostman HD, Brizel DM, Layfield L, Charles HC, Dewhirst MW, Prescott DM, Friedman HS, Harrelson JM, Scully SP, Coleman RE. Monitoring of neoadjuvant therapy response of soft tissue and musculoskeletal sarcoma using fluorine- 18-FDG PET. J Nucl Med
12. Kalff V, Hicks RJ, MacManus MP, Binns DS, McKenzie AF, Ware RE, Hogg A, Ball DL. Clinical impact of 18F fluorodeoxyglucose positron emission tomography in patients with non-small-cell lung cancer: a prospective study. J Clin Oncol
13. Kern KA, Brunetti A, Norton JA, Chang AE, Malawer M, Lack E, Finn RD, Rosenberg SA, Larson SM. Metabolic imaging of human extremity musculoskeletal tumors by PET. J Nucl Med
. 1988;29: 181-186.
14. Kole AC, Nieweg OE, Hoekstra HJ, van Horn JR, Koops HS, Vaalburg W. Fluorine-18-fluorodeoxyglucose assessment of glucose metabolism in bone tumors. J Nucl Med
15. Kole AC, Nieweg OE, van Ginkel RJ, Pruim J, Hoekstra HJ, Paans AM, Vaalburg W, Koops HS. Detection of local recurrence of soft tissue sarcoma with positron emission tomography using [18F] fluorodeoxyglucose. Ann Surg Oncol
16. Lucas JD, O'Doherty MJ, Wong JC, Bingham JB, McKee PH, Fletcher CD, Smith MA. Evaluation of fluorodeoxyglucose positron emission tomography in the management of soft tissue sarcomas. J Bone Joint Surg Br
17. Malawer MM, Link MP, Donaldson SS. Sarcomas of bone. In De- Vita VT, Hellman S, Rosenburg SA, eds. Cancer: Principles and Practice of Oncology
. Philadelphia: JB Lippincott Company; 1993: 1509-1566.
18. Malawer MM, Sugarbaker P. Musculoskeletal cancer surgery: treatment of sarcomas and allied diseases. Dordrecht, Netherlands: Kluwar Academic Publishers; 2001:4.
19. Simon MA, Kirchner PT. Scintigraphic evaluation of primary bone tumors: comparison of technetium-99 mm phosphate and gallium citrate imaging. J Bone Joint Surg Am