Radionuclide imaging of soft tissue neoplasms

Radionuclide Imaging of Soft Tissue Neoplasms Felix S. Chew, Terry M. Hudson, and William F. Enneking T w o classes of radiopharmaceuticals may be used for imaging tumors of the musculoskeletal system. The first is comprised of soft tissue or tumorspecific agents such as gallium-67, bleomycin, and radionuclide-labeled antibodies, which may be useful for detecting and localizing these tumors. The other class of tracer is comprised of those with avidity for bone. The ~SmTc-labeled-phosphate skeletal imaging compounds have been found to localize in a variety of soft tissue lesions, including benign and malignant tumors. In 1972, Enneking began to include bone scans in the preoperative evaluation of soft tissue masses. Later, he and his associates reported t h a t these scans w e r e useful in planning operative treat-

ment of sarcomas by detecting involvement of bone by the tumors. Nearly all malignant soft tissue tumors take up bone-seeking radiopharmaceuticals, and bone involvement was indicated in two-thirds of the scans we reviewed. About half of benign soft tissue lesions had normal scans, but the other half showed uptake w i t h i n the lesion and a f e w also showed bone involvement. Careful, thorough imaging technique is essential to proper evaluation. Multiple, high-resolution static gamma camera images in different projections are necessary to adequately demonstrate the presence or absence of soft tissue abnormality and to define the precise relationship of the tumor to the adjacent bone.

WO T Y P E S of radiopharmaceuticals may be used for imaging soft tissue tumors of the musculoskeletal system. First are those agents intended specifically for localization in soft tissue tumors, such as gallium and bleomycin. Second are the technetium-99m-labeled phosphate compounds intended for imaging of the skeleton. After the introduction of these bone-seeking agents, reports of localization in tumors and other lesions of soft tissues began to appear, and now a large number of neoplastic and inflammatory lesions of the soft tissues are known to accumulate the skeletal tracers. Our review of a large number of bone scans has shown that such soft tissue uptake is not only common, it is typical of most malignant and many benign soft tissue tumors and can be of value in diagnosis and in surgical treatment planning.

gallium scans in 4 of 8 sites of soft tissue sarcomas, but only 2 positive scans of 8 sites of bone metastases from these lesions. A report by Kaufman et al. 3 included gallium scans of 27 soft tissue sarcomas, of which 25 were positive. This included 10 tumors that arose in the extremities and limb girdles. Nine of the latter scans were positive. Bitran et al. 4 reported a series of gallium scans that included 29 soft tissue sarcomas. The accuracy of the scans varied with the specific diagnosis. In one group of 16 tumors, including malignant schwanoma, rhabdomyosarcoma, and undifferentiated sarcoma, 29 of 31 sites of tumor were positive on the scan: a 94% sensitivity. However, in another group of 13 tumors, including liposarcoma, leiomyosarcoma, malignant fibrous histiocytoma ( M F H ) , and synovioma, only 5 of 16 sites of tumor were positive: a sensitivity of only 31%. In 1978, Teates et al. 5 reviewed the literature dealing with 67Ga-citrate scanning. Fourteen of 15 scans were positive in patients with M F H , rhabdomyosarcoma, or liposarcoma. They found 59 positive gallium scans in 65 patients with sarcomas, but many of these were thoracic or abdominal tumors, not musculoskeletal soft tissue tumors. At least some of the patients they included were among those reported by Lepanto and Kaufman. Pinsky and Henkin 6 discussed the use of 67Gacitrate in scanning soft tissue tumors, the technique of scanning, and the interpretation of normal scans, and listed the nontumor causes of increased uptake, including abscess, sarcoid, rheumatoid arthritis, Paget's disease, fracture, surgery, and others. They concluded that gallium is still "the best available isotope" for

T

SOFT TISSUE IMAGING AGENTS

Gallium-67-Citrate Hayes ~ has recently reviewed the history of the development of 67Ga citrate, including its evolution into an agent used specifically to study soft tissue malignancies and inflammatory processes. Lepanto et al. 2 reported positive From the Departments of Radiology and Orthopedic Surgery and the W. Thaxton Springfi'eld Center for Orthopedic Study and Research, University of Florida Medical Center, Gainesville, Fla. Reprint requests should be addressed to Dr. Terry M. Hudson, Box J-374 J.H.M.H.C., Department of Radiology, University of Florida, Gainesville, Fla. 32610. 9 1981 by Grune & Stratton, Inc. 0001-2998/81/1104-0005502.0(9/0 266

Seminars in Nuclear Medicine, Vol. XI, No. 4 (October), 1981

IMAGING OF SOFT TISSUE NEOPLASMS

tumor scanning and listed the wide variety of epithelial and nonepithelial tumors that show avidity for 67Ga-citrate. Larson 7 reviewed the radiopharmacy of 67Ga-citrate and the various possible mechanisms of its localization in tumors, including the possible important role of transferrin. However, the exact mechanism is not yet known with certainty.

Other Tumor-Scanning Agents Bleomycin labeled with various radionuclides has been used as a tumor-specific scanning agent. Goodwin and Mears 8 reviewed the chemical and biologic properties of this agent, including the relative merits of the various radionuclide labels. Mori et al. 9 found that 99mTcbleomycin scans detected 80% of various malignant tumors, including 7 of 8 "cancer and sarcoma" of the extremities, and they discussed the properties of bleomycin as a prototype radiopharmaceutical for imaging tumors. Immunologic mechanisms are also potentially valuable means of imaging tumors. Spar t~ discussed the labeling of tumor-specific antibodies with detectable radionuclides, but this technique is not presently clinically practicable. SKELETAL IMAGING AGENTS

Technetium-99m-labeled phosphate compounds, intended for skeletal imaging, are taken up by a wide variety of soft tissue abnormalities, including various benign and malignant tumors, cerebral infarction, myocardial infarction, soft tissue calcification or ossification, inflammatory lesions such as abscesses, surgical wounds, hematomas, and others. ~ Most of the soft tissue tumors that have been reported to accumulate these agents have been malignant, although Nolan j: reported incidental discovery of a benign neurofibroma on a 99mTc-diphosphonate scan. Pearlman ~3 reported two 99mTc-diphosphonate and one 99mTc-pertechnetate scan of liposarcoma including dynamic flow studies and static images. He found the concentration of the radiopharmaceutical in each tumor to be roughly proportional to blood flow. Blatt et al. ~4 reported imaging of a liposarcoma with 99mTc-pyrophosphate and felt that hypervascularity and microscopic calcification in the tumor contributed to the radionuclide uptake. In 1972, Enneking 15began to include radionuclide bone scans in the preoperative radiologic

267

staging of soft tissue masses. We reviewed most of these scans and found that they helped differentiate between benign lipomas and liposarcomas, 16 and we also reported uptake in some benign infiltrating angiolipomas. ~7

Bone Scanning in Planning Surgical Treatment of Soft Tissue Tumors Bone metastases from primary soft tissue sarcomas are unusual. Felix et al. ~8 detected no metastases on the radionuclide bone scans of 59 patients with sarcomas, and we found only 5 patients with metastases on the bone scans of 80 sarcoma patients. Since bone metastases are unusual, the value of the radionuclide bone scan in preoperative staging lies in the evaluation of the relationship of the primary tumor to adjacent bone. Enneking and his associates have shown that the surgical margin required for adequate treatment of a soft tissue sarcoma is dictated by the histologic grade of the tumor and that the operative procedure required to achieve the desired margin is dictated by the anatomic location and extent of the tumor. ~9-2~ They assess the anatomic extent by means of clinical history, physical examination, plain radiography, arteriography, computed tomography, and radionuclide bone scanning. When the presence of a palpable soft tissue mass suggests the possibility of malignancy, complete anatomic staging should be done prior to biopsy, because biopsy before staging can contaminate tissue planes with tumor and confuse subsequent radiologic studies. Contamination occurs when hematoma dissects along fascial planes far beyond the margins of the biopsy incision. This hematoma may carry tumor cells with it and sarcomas, in contrast to carcinomas, readily implant in soft tissue. Consequently, widespread tumor seeding may necessitate more extensive surgical resection or radiation therapy than would have otherwise been required. ~9'2~Tumor evaluation is also complicated by the hematoma, edema, scarring, and neovascularity that follow biopsy so that it may be very difficult to determine the anatomic extent of the tumor by physical examination, arteriography, or computed tomography. ~9'2~The same is true of the radionuclide scan, since it is well recognized that surgical wounds alone can act as foci of hyperconcentration. ~ Adequate surgical resection of aggressive

268

CHEW, HUDSON, AND ENNEKING

tumors requires removal of all involved structures. Soft tissue tumors elicit a hyperemic, reactive response in surrounding soft tissues. Within this surrounding reactive tissue (pseudocapsule), microscopic tumor spread can be found pathologically. ~9"2~ Therefore, when either the tumor itself, or its reactive pseudocapsule is in contact with bone, the bone is considered involved and must be resected. Enneking et al. 22 have demonstrated that on the radionuclide bone scan, increased uptake in bone adjacent to a soft tissue sarcoma indicates bone involvement. Such involvement may also be present when hyperconcentration in the soft tissue lesion itself is contiguous with the bone and cannot be separated, even on appropriate multiple scan views. Therefore, the bone scan may be useful when knowledge of the tumor's relationship to bone is critical to planning the appropriate resection. TECHNIQUE OF SCANNING

The imaging technique used in this series varied somewhat. The radiopharmaceutical was 99mTc-pyrophosphate in a few instances, but 99roTe-methylene diphosphonate was used in the vast majority. The usual adult dose was 15-20 mCi (550-740 MBq) and the interval between injection and the beginning of scanning was 2-3 hr. A proportionately smaller dose was used in children. Early in our series, between 1972 and 1976, 31 scans were obtained with an Ohio Nuclear Series 84 rectilinear scanner. After 1976, 76 scans were obtained with an Ohio Nuclear Series 100 gamma camera in a whole body imaging configuration. In 56 of these patients, additional static gamma camera views with a high resolution collimator were obtained, each including approximately 200,000 counts. Six scans were performed at other hospitals. It was quite evident in reviewing the scans that the rectilinear images were vastly inferior to the gamma camera views. Faintly perceptible abnormalities on the whole body camera images frequently were clear and distinct abnormalities on the high-resolution static camera views. In addition, determination of the relationship of the soft tissue activity to the nearby bone required multiple camera views (see Figs. 2 and 5). An adequate study often included not only anterior, posterior, and lateral views, but frequently individually tailored oblique views to demonstrate

the exact tangential relationship of the tumor to the adjacent bone (see Fig. 2). Approximately 200,000 counts per static gammera image seemed appropriate, although occasionally, images were not optimal and perhaps more counts should have been collected. The number of counts should be appropriately adjusted to the filmed image size used in each individual department. Comparison gamma camera views of the contralateral normal side are frequently essential to decide whether there is abnormal soft tissue activity or to decide whether there is focal hyperactivity in the adjacent bone. Preset imaging times rather than equal numbers of counts per image should be used for comparison views to avoid obscuring slight differences in activity. It is essential that the scan be supervised by a physician familiar with the clinical problem. When a patient with a soft tissue tumor is referred for radionuclide imaging, the supervising physician should be aware of the clinical findings and abnormalities demonstrated by other radiographic studies. When soft tissue uptake is found incidentally on a whole body bone scan, additional camera views and additional clinical and radiographic information should be obtained. PATTERNS OF SCANS Patterns of Scan A b n o r m a l i t y

At the University of Florida Medical Center, more than 240 patients with soft tissue masses were evaluated with bone scans using 99roTelabeled phosphate compounds. We reviewed these and analyzed in detail 113 patients for whom complete material was available, including adequate scan images and surgical and pathologic correlation of the anatomic extent of the tumor. All of these patients presented with clinically evident soft tissue masses for which sarcoma was a reasonable diagnostic possibility. The scan results were variable. Some scans were normal, but most disclosed hyperconcentration of the tracer in the soft tissue lesion and sometimes in the adjacent bone. Some of the masses had already been biopsied. Although the biopsy wound alone might produce an area of increased uptake on the scan, we found no difference in comparing the scans of patients with and without prior biopsy. That is, those types of tumors that

IMAGING OF SOFT TISSUE NEOPLASMS

were positive on scan did not show greater uptake on scans following incisional biopsy than on scans of patients whose lesions had not been biopsied. Normal Scans

Only 29 of the 113 scans were entirely normal. Only one was a malignant tumor: a soft tissue chondrosarcoma arising in the buttocks. All the other lesions were benign: lipoma, 11; infiltrating angiolipoma, 3; fibromatosis and ganglion, 2 each; neurofibroma, hematoma, tumoral calcinosis, hamartoma, abscess, tuberculous granuloma, extraabdominal desmoid and synovitis, 1 each; no pathologic diagnosis, 3 patients.

269

ranged from ill defined and barely perceptible to well marginated and vivid. There was usually very intense radionuclide uptake in areas of radiographically evident calcification (Fig. 1). Apart from this relationship, we could find no absolute correlation between the intensity of radiopharmaceutical uptake and the histogenesis, size, or vascularity of the tumor, or whether there had been previous incisional biopsy. However, when angiograms were available, hypervascularity was present in most of the lesions, which demonstrated radionuclide hyperconcentration. There were five patterns of radionuclide abnormality.

( l ) Isolated Soft Tissue Hyperconcentration Abnormal Scans

Eighty-four scans revealed abnormal radionuclide hyperconcentration in or around the soft tissue lesion. The intensity of hyperconcentration

In 36 patients, there was hyperconcentration of radionuclide activity only within the soft tissue lesion itself (Figs. 2 and 3). Twelve tumors were malignant: malignant fibrous histiocytoma,

Fig, 1. Benign lipoma. (A) Lateral radiograph of distal thigh shows mass with lucency of fat, containing areas of dense calcification. (B) Lateral camera view of scan shows uptake in calcified areas. (Reproduced by permission of Radiology.TM)

270

CHEW, HUDSON, AND ENNEKING

Fig. 2. Malignant schwannoma. (A) Posterior whole body image shows an area of ill defined increased uptake in t h e distal left thigh (arrow). (B) Posterior, high-resolution static camera v i e w shows abnormal uptake much more distinctly. (C) Left posterior oblique camera v i e w was required to demonstrate separation of soft tissue uptake from normal activity in femur.

IMAGING OF SOFT TISSUE NEOPLASMS

271

parts, 1 each. Seven lesions were benign: fibromatosis, 5; and abscess and ganglion, 1 each. We have already mentioned that in malignant soft tissue tumors, hyperconcentration in adjacent bone indicates involvement of that bone by the tumor or by reactive tissue around it. 22 In the present series, all seven of the patients with benign lesions and hyperconcentration in adjacent bone also had involvement of the bone demonstrated pathologically or surgically. There was erosion of underlying bone by all five cases of fibromatosis and by the ganglion, and reactive periosteal bone formation adjacent to the soft tissue abscess.

(3) Hypereoncentration in Soft Tissue Contiguous With Normal Bone

Fig. 3. Benign infiltrating angiolipoma in medial right thigh shows moderately intense activity (arrow), (Reproduced by permission of the American Journal of Roentgenology. ~~)

(MFH), 4; liposarcoma and undifferentiated sarcoma, 2 each; synovioma, chondrosarcoma of soft parts, schwannoma, rhabdomyosarcoma, and hemangiopericytoma, 1 each. Twenty-two of the lesions were benign: infiltrating angiolipoma, 6; hematoma and lipoma, 3 each; neurofibroma and myxoma, 2 each; neurilemoma, lymphangiomyoma, myositis ossificans, hamartoma, synovitis, and aggressive fibromatosis, 1 each. No lesion was found in one patient.

(2) Hyperconcentration in Soft Tissue and Adjacent Bone In 31 patients, there was hyperconcentration of the radionuclide activity, not only within the soft tissue lesion itself, but also focally in the immediately adjacent bone, with or without radiographically evident bone abnormality (Figs. 4 and 5). Twenty-four of these lesions were malignant: MFH, 7; liposarcoma, 5; fibrosarcoma, rhabdomyosarcoma, synovioma, and undifferentiated sarcoma, 2 each; malignant schwannoma, hemangiopericytoma, Ewing's sarcoma of soft parts, and osteosarcoma of soft

In eight patients, there was no abnormally increased activity in bone, but the hyperconcentration of radionuclide within the soft tissue lesion appeared contiguous with the normal activity in adjacent bone on all available views. Adequate demonstration of this situation required a gamma camera view showing the tangential relationship of the soft tissue lesion to the bone. All of these lesions were malignant: liposarcoma and neuroblastoma, two each; MFH, fibrosarcoma, histiocytic lymphoma, and leiomyosarcoma, one each.

(4) Unlocalizable Hyperconcentration In seven patients, there was abnormal radionuclide activity in the vicinity of the soft tissue lesion, but for various reasons we could not clearly localize this activity to soft tissue or bone. in some of these, additional camera views might have clarified the situation, but some lesions in the limb girdles simply could not be imaged optimally (Fig. 6). Six of these patients had malignant lesions: MFH, three; neural sarcoma, angiosarcoma, and synovioma, one each. One patient had a benign hemangioma.

(5) Extended Uptake In two instances, the only scintigraphic abnormality was extended uptake 23 in the bones of the lesion-bearing limb. One patient had neurofibromatosis and the other had a ganglion. DIFFICULTIES IN INTERPRETATION

Incidental causes of radionuclide hyperconcentration in the tissues near a soft tissue mass

272

CHEW, HUDSON, AND ENNEKING

Fig. 4. Liposarcoma of thigh stimulating visible periosteal reaction. (A and B) Anteroposterior and lateral radiographs show soft tissue mass (curved arrows) and thick layer of pariosteal n e w bone on femur (straight arrow). (C and D) Anterior and lateral camera scintigrams show intense uptake within tumor (straight arrow) and " s t r i p e " of intensely increased bone activity corresponding to n e w bone formation (curved arrow). Since the plain radiographs revealed the bone abnormality, the scan did not add new information.

Fig. 5. Recurrent malignant fibrous histiocytome. (A) Lateral radiograph shows soft tissue mass (arrow) anterior to n o r m a l tibia. (B and C) A n t e r i o r and lateral camera images disclose increased uptake in the tibia adjacent to the t u m o r (arrow). The t w o right-angle views make clear that the bone itself is abnormal and therefore involved by tumor. The appearance is not due simply to superimposition of soft tissue activity over t h e tibia.

274

CHEW, HUDSON, AND ENNEKING

Fig. 6. Malignant fibrous histiocytoma of shoulder girdle. Posterior camera image shows intense, well defined activity (arrow) but w e could not be sure whether it involved only soft tissue or the scapula as well. The anatomical location made it impossible to obtain images to solve the problem.

will cause difficulties in interpretation and may even invalidate the results of the scan. We found various benign skeletal and articular lesions causing radionuclide hyperconcentration in about 10% of our patients, including arthritis and old trauma. When a soft tissue mass is located near such a skeletal abnormality or near an active epiphyseal growth plate, it is virtually impossible to determine whether there is truly hyperconcentration in the adjacent bone. The cause of such incidental skeletal uptake can usually be identified after careful clinical and radiologic evaluation. The extended or augmented uptake pattern is another potential cause of spurious hyperconcentration in adjacent bone when there is no real involvement of the bone by the tumor. In this pattern, diffusely increased radionuclide activity is present along the long bones of the lesionbearing extremity, usually with epiphyseal accentuation. 23'24 Plain radiographs may be normal or may demonstrate osteopenia. Extended uptake can usually be distinguished from the appearance of true bone involvement, since the increased uptake is present in the entire limb and is not localized to bone immediately adjacent to the soft tissue tumor. Radionuclide activity associated with vascular

calcifications, 1~ especially the femoral arteries, can be confusing when the soft tissue mass is located near the vessel. Recognition of the linear pattern of radionuclide uptake and the typical calcification visible on the plain radiograph will usually resolve the difficulty. Since soft tissue trauma, including surgical incision, can produce focal uptake on the scan, confusion can result if the patient has undergone recent biopsy. Soft tissue activity may then be due to the surgery rather than the tumor itself. However, we found that the presence of a surgical wound did not seem to increase the intensity of the uptake in lesions that typically demonstrate soft tissue activity, and furthermore, even after incomplete surgical excision, the scan is still valuable in demonstrating tumor involvement of adjacent bone. 22 This is true because of the assumption that any tissue that was entered during biopsy is potentially contaminated by tumor cells and must therefore be removed along with all gross tumor during definitive surgical resection. J9-22 SCINTIGRAPHIC PATHOLOGIC CORRELATIONS

In our series, all but one of the patients with normal scans had benign processes or no identifiable lesion at all. These lesions typically did not

IMAGING OF SOFT TISSUE NEOPLASMS

275

involve the underlying bone, were uncalcified on plain radiographs, and did not display hypervascularity on angiograms. However, many benign lesions did demonstrate radionuclide uptake. Isolated soft tissue hyperconcentration occurs in a wide variety of benign and malignant tumors. The uptake of skeletal tracers in areas of soft tissue calcification or ossification is well known, ~ and we noted intense uptake in the calcified areas of various lesions (Fig. 1). The mechanism of radionuclide concentration in noncalcified lesions is not well understood. Hypervascularity has been related to abnormal uptake in soft tissue lesions. 13'14'16J7 We could not demonstrate a precise correlation between the intensity of radionuclide hyperconcentration and the degree of hypervascularity, although most of the lesions we found positive on scan were either hypervascular on an angiogram or were of a histogenesis commonly associated with hypervascularity. However, some vascular lesions were negative (angiolipoma, neurofibroma) and some relatively hypovascular, uncalcified lesions were positive (fibromatosis, ganglion, hematoma). Hyperconcentration in adjacent bone occurs

in patients with malignant or benign lesions involving the bone. In soft tissue sarcomas, the pathologic explanation is stimulation of periosteal reaction due to direct involvement of the bone by either the tumor itself or by the reactive mesenchymal tissue about it. When plain radiographs reveal erosion of the bone or periosteal new bone, the radionuclide scan is superfluous (Fig. 2). However, the scan can reveal bone involvement when the bone appears normal on the roentgenograms (Fig. 3). Similarly, soft tissue abscesses may evoke reaction in the adjacent bone affected by the hyperemic inflammatory tissue surrounding the infection. Benign lesions, on the other hand, do not typically elicit a hyperemic reactive response in tissue beyond the actual margins of the lesion. Those benign lesions that stimulated increased scan uptake in adjacent bone pathologically showed actual cortical erosion directly by the tumor itself. ACKNOWLEDGMENT

Dr. Clyde M. Williams, Professor and Chairman of the Department of Radiology and Head of Nuclear Medicine, supervised most of the bone scans and made them freely available to us.

REFERENCES

1. Hayes RL: The medical use of gallium radionuclides: A brief history with some comments. Semin Nucl Med 8:183 191, 1978 2. Lepanto PB, Rosenstock J, Littman P, et al: Gallium67 scans in children with solid tumors. Am J Roentgenol 126:179-186, 1976 3. Kaufman JH, Cedermark BJ, Parthasarathy KL, et al: The value of 67Ga scintigraphy in soft-tissue sarcoma and chondrosarcoma. Radiology 123:131-134, 1977 4. Bitran JD, Bekerman C, Golomb HM, et al: Scintigraphic evaluation of sarcomata in children and adults by Ga 67 citrate. Cancer 42:1760-1765, 1978 5. Teates CD, Bray ST, Williamson BRJ: Tumor detection with 67Ga-citrate: A literature survey (1970-1978). Clin Nucl Med 3:456-460, 1978 6. Pinsky SM, Henkin RE: Gallium-67 tumor scanning. Semin Nucl Med 6:397-408, 1976 7. Larson SM: Mechanisms of localization of gallium-67 in tumors. Semin Nuc[ Med 8:193 203, 1978 8. Goodwin DA, Meares CF: Radiolabeled antitumor agents. Semin Nucl Med 6:389-396, 1976 9. Mori T, Hamamoto K, Onoyama Y, et al: Tumor imaging after administration of 99mTc-labeled bleomycin. J Nucl Med 16:414-422, 1975 10. Spar NG: An immunologic approach to tumor imaging. Semin Nucl Med 6:379-387, 1976

11. Citrin DL, McKillop JH: Atlas of Technetium Bone Scans. Philadelphia, Saunders, pp 43-65 12. Nolan NG: Intense uptake of 99mTc-diphosphonate by an extraosseous neurofibroma. J Nucl Med 15:1207-1208, 1974 13. Pearlman AW: Preoperative evaluation of liposarcoma by nuclear imaging. Clin Nucl Med 2:47 51, 1977 14. Blatt C J, Hayt DB, Desai M, et al: Soft-tissue sarcoma: Imaged with technetium-99m pyrophosphate. NY State JMed77:2118 2119, 1977 15. Enneking WF: Clinical Musculoskeletal Pathology. Gainesville Fla, Storter, 1977, pp 377-396 16. Chew FS, Hudson TM: Radionuclide imaging of lipoma and liposarcoma. Radiology 136:741-745, 1980 17. Chew FS, Hudson TM: Radiology of infiltrating angiolipoma. Am J Roentgenol 135:781-787, 1980 18. Felix EL, Sindelar WF, Bagley DH, et al: The use of bone and brain scans as screening procedures in patients with malignant lesions. Surg Gynecol Obstet 141:867-869, 1975 t9. Simon MA, Enneking WF: The management of soft tissue sarcomas of the extremities. J Bone Joint Surg 58:317 327, 1976 20. Simon MA, Spanier SS, Enneking WF: Management of adult soft-tissue sarcomas of the extremities, in Nyhus LM (ed): Surgery Annual. New York, Appleton Century Crofts, 1979 21. Enneking WF, Spanier SS, Goodman MA: A system

276

for the surgical staging of muscuioskeletal sarcoma. Clin Orthoped 153:106-120, 1980 22. Enneking WF, Chew FS, Springfield DS, et al: The role of radionuclide bone-scanning in determining the resectability of soft-tissue sarcomas. J Bone Joint Surg 63:249 257, 1981

CHEW, HUDSON, AND ENNEKING

23. Thrall JH, Ghaed N, Geslien GE, et al: Pitfalls in Tc-99m polyphosphate skeletal imaging. Am J Roentgenol 121:739-747, 1974 24. Goldman AB, Braunstein P: Augmented radioactivity on bone scans of limbs bearing osteosarcomas. J Nucl Med 16:423, 1975