Imaging of Bone Sarcomas

JN054_Jrnl_50408Cummi.qxd 4/12/07 12:23 AM Page 438

438

Original Article

Judd E. Cummings, MD; J. Andrew Ellzey, MD; and Robert K. Heck, MD, Memphis, Tennessee

Key Words Diagnostic Evaluation Bone sarcoma, imaging, staging Patient evaluation begins with a thorough history and physical examination. Most commonly, patients present Abstract with pain at the affected site. The pain usually does not Identification, staging, and treatment of bone sarcomas rely on both improve and may progress to become independent of clinical and imaging evaluations. Although conventional radiography remains the primary imaging modality for characterizing bone activity. Ultimately, the patient may experience pain at tumors, bone scintigraphy, computed tomography, magnetic rest or at night. resonance imaging, and positron emission tomography can each Although some tumors show a sex predilection (fe- add information for staging and treatment planning. (JNCCN male predominance with giant cell tumor and parosteal 2007;5:438–447) osteosarcoma), this rarely is useful for diagnosis. Similarly, race is of little diagnostic benefit except with Ewing’s sar- Although much less prevalent than other forms of can- coma, which is extremely rare in people of African de- cer, bone sarcomas pose a dilemma in both diagnosis and scent. In contrast, the patient’s age is very helpful in treatment. Fortunately, over the past several decades, the directing the differential diagnosis of a suspected bone advancement and refinement of modalities aimed at treat- tumor.14,15 Examples include primary osteosarcoma, which ing bone sarcomas, such as adjuvant chemotherapy and usually occurs between the ages of 10 and 25 years; Paget’s limb-sparing surgical techniques, have resulted in better osteosarcoma, between 55 and 80 years; Ewing’s sarcoma, 1–13 outcomes for many patients. Appropriate care of pa- between 5 and 25 years; primary chondrosarcoma, be- tients with bone sarcomas is heavily dependent on a tween 30 and 60 years; unicameral and aneurysmal bone prompt and thorough initial evaluation, and a working cysts, younger than 20 years; chondroblastoma, younger knowledge of the various imaging techniques to evaluate than 20 years; giant cell tumor, between 20 and 40 years; bone tumors is paramount for any physician who may en- and metastasis and multiple myeloma, older than 40 counter one of these patients early in the disease course. years.16 Various imaging studies are used in diagnosing, stag- Physical examination should include evaluation of ing, and monitoring bone sarcomas. Conventional radi- the patient’s general health in addition to the affected ography, bone scintigraphy, computed tomography (CT), part. Constitutional symptoms, fever, weight loss, and magnetic resonance imaging (MRI), and positron emis- night pain should be documented. Any mass should be sion tomography (PET) each have strengths and limita- measured and inspected for consistency, mobility, pain, tions. This article describes how these modalities are temperature, fluctuance, and skin changes. A thorough applied to the evaluation and treatment of suspected bone neurovascular examination is recorded and potential sites sarcomas. of lymph node metastasis are palpated. Imaging studies are then used to formulate a working diagnosis and appropri- ately stage the lesion. If these studies, along with the clin- From the University of Tennessee-Campbell Clinic, Department of Orthopaedic Surgery, Memphis, Tennessee. ical information, suggest a malignant bone tumor, the Submitted October 24, 2006; accepted for publication November patient should be referred to a musculoskeletal oncology 13, 2006. The authors have no financial interest, arrangement, or affiliation center without additional tests or biopsy. For several rea- with the manufacturers of any products discussed in the article or their competitors. sons, bone tumors must be completely evaluated before Correspondence: Robert K. Heck, MD, 1211 Union Avenue, Suite proceeding with a biopsy. A thorough evaluation helps 510, Memphis, TN 38104. E-mail: [email protected] narrow the differential diagnosis and leads to a more

Original Article 439

Imaging of Bone Sarcomas

accurate pathologic diagnosis. The type of biopsy performed and placement of the biopsy incision are de- termined by the extent and potential resectability of the lesion. Additionally, the accuracy of imaging studies such as CT, MRI, or bone scan may be adversely affected by postoperative changes from a premature biopsy.17–22 Biopsy should be performed by a surgeon familiar with musculoskeletal oncology techniques, preferably by one who will perform any definitive pro- cedures. Similarly, if a radiologist is to perform the biopsy, the treating surgeon should be consulted. Conventional Radiography Since the discovery of the x-ray in 1895 by Wilhem Conrad Roentgen, continuous innovations have oc- curred in the application of radiographic imaging to study human disease.23,24 Today, it remains the most widely used imaging modality in medicine and contin- ues to be the gold standard for creating an accurate differential diagnosis of suspected bone lesions.23–27

The appearance of a plain radiograph is a function of Figure 1 Anteroposterior radiograph of the left elbow of a 3-year-old x-ray beam attenuation as it passes through anatomic patient shows a thick, continuous periosteal reaction consistent with structures. The extent of beam attenuation is a func- a benign diagnosis, such as infection, healing fracture, or benign bone tion of the energy of the x-ray beam and the tissue’s tumor. density. Tissues with high attenuation, such as bone, periosteal reaction is frequently associated with appear relatively white, whereas fat or lung tissue al- malignant diagnoses such as osteosarcoma and Ewing’s lows most of the beam to pass through to the film, sarcoma (Figures 2 and 3). producing relatively dark areas.26,27 The appearance of a tumor’s margin is extremely The radiograph produced by the contrasting den- important. In general, most benign bone lesions have sity of human tissues produces a wealth of informa- well-defined margins with a rim of reactive bone sur- tion when analyzed carefully. Plain radiographs should rounding the tumor (Figure 4). Malignant bone le- always be the initial imaging modality for evaluating sions, on the other hand, have poorly defined margins patients with a suspected bone lesion.26,28 Frequently, with a wide zone of transition to normal bone and the patient’s age combined with findings on plain ra- show patchy cortical destruction with early soft-tis- diographs provide enough information to yield an ac- sue extension (Figure 5).32 curate diagnosis.15,26,29 Physicians should consider the Additionally, lesion location must be considered. presence of bone destruction, tumor matrix, periosteal Chondroblastoma (younger patients), giant cell tumor or endosteal reaction, lesion margins, associated soft- (adults), and clear cell chondrosarcoma show a tissue changes, and lesion location.26 The appearance predilection for the epiphyses of bones. In contrast, of the tumor matrix can suggest a particular diagno- Ewing’s sarcoma, histiocytosis, and adamantinoma are sis; infarcts produce a “smoke in the chimney” appear- found more commonly in the diaphyses of long ance, fibrous dysplasia yields a “ground glass” bones.15,22 In the spine, lesions are considered based on appearance in bone,15,30 chondroid lesions often show their location within the vertebral body or the poste- calcified “rings and arcs,”31 and osteosarcoma and os- rior elements. Adults with a lesion of the vertebral teoblastoma frequently show matrix ossification. body usually have metastasis, myeloma, or a Periosteal reactions have been termed solid, spiculated, hemangioma. A vertebral body lesion in a young Codman triangle, or unorganized.15,24 In general, thick patient typically represents histiocytosis, whereas continuous periosteal reaction (Figure 1) is associated lesions in the posterior elements in younger patients with benign entities such as stress fracture, histio- most commonly are osteoid osteoma, osteoblastoma, or cytosis, and infection, whereas thin, discontinuous aneurysmal bone cyst.16 Collectively, this information

440 Original Article

Cummings et al.

Figure 2 Anteroposterior (A) and lateral (B) radiographs of the right distal femur of a 7-year-old patient with osteosarcoma show an ill-de- Figure 3 Anteroposterior radiograph of the right proximal femur of a fined bone-forming lesion with periosteal reaction in the form of a 16-year-old patient with Ewing’s sarcoma shows thin, wispy, “onion Codman triangle. skin” periosteal reaction.

should allow an accurate differential diagnosis that can ing unless the lesion is purely lytic, such as multiple help direct further workup and treatment.15,30 myeloma or possible renal cell metastasis, which can 33 Bone Scintigraphy produce false-negative studies. Bone scans detect areas of increased bone metabo- CT lism. A radionuclide, most commonly technetium The development of the CT scanner is credited to (Tc)-99m–labeled diphosphonate, is administered in- Godfrey N. Hounsfield in 1973.38 An advancement travenously and allowed to accumulate in bone. As the of routine radiography, the image produced is still a radiotracer decays, gamma radiation is emitted and function of the differential absorption of x-ray beams detected with a camera scanning the patient. Increased as they pass through human tissue. With CT, the uptake is shown in areas of increased vascularity or x-ray beam is collimated into a narrow beam that bone metabolism. Whole-body bone scans can detect passes through the patient in thin slices from multi- the presence of skip metastases, or distant sites of bone ple angles. The beam is absorbed by highly sensitive involvement important for staging. They also are used detectors (single or multiple) capable of identifying for postoperative evaluation to rule out local recur- subtle differences in tissue density. The scanner rence or late bone metastasis.33–35 computes a tomographic (single slice in 3 dimensions) Because plain radiography requires up to a 50% loss image that can be reformatted for visualization in dif- of mineralization for detecting a bone lesion, bone ferent planes (sagittal or coronal). Images can be scintigraphy sometimes can identify the presence of an enhanced by administering contrast (oral or intraosseous abnormality before it is visible on plain films venous), giving the CT excellent spatial and contrast (Figure 6).31 Unfortunately, radionuclide bone scan- resolution that makes it particularly helpful for ning has very limited potential to differentiate benign evaluating lesions in the axial skeleton, such as the from malignant osseous or chondroid lesions, because pelvis or spine, where complex anatomy and overlap many benign bone tumors show increased radiotracer of anatomic structures can make 2-dimensional analy- uptake.36,37 A positive bone scan, therefore, indicates sis difficult (Figure 7).14,27,39–41 Cross-sectional and re- the presence of a lesion that requires further evalua- formatted 3-dimensional images provide valuable tion. A negative bone scan, however, is very reassur- characterizations of bone abnormalities and disease

Original Article 441

Imaging of Bone Sarcomas

Figure 4 Anteroposterior radiograph of the right proximal humerus of Figure 5 Anteroposterior radiograph of the right distal femur of a an 8-year-old patient with a healing pathologic fracture secondary to 12-year-old patient with malignant fibrous histiocytoma of bone shows unicameral bone cyst shows benign imaging features: well-defined mar- aggressive imaging features: poorly defined margins with extensive bone gins, a narrow zone of transition, and a rim of sclerotic reactive bone. destruction. processes. CT is used to show the type or presence of to tumor staging and the planning of limb salvage cortical bone destruction and presence of matrix min- surgery.31 eralization.17,30,35 A sclerotic rim on CT imaging often MRI uses the magnetic properties of atomic nuclei suggests a benign process, and the presence of fluid (protons) to generate and detect signals that can be levels characterizes cystic structures.30,34 Finally, CT converted into a grayscale image. The most abundant imaging may confirm the presence of a pathologic of these atoms within human tissue is hydrogen-1, which fracture or the characteristic nidus of an osteoid os- is primarily responsible for the nuclear magnetic reso- teoma. nance (NMR) signal used to generate a traditional MRI. Although MRI has replaced CT for local staging Some nuclei, such as sodium-23, phosphorous-31, and of bone tumors in most cases,34,40 CT scanning remains fluorine-19, are capable of generating an NMR signal a vital part of this process because it is the preferred but are several orders of magnitude less abundant within test for evaluating the presence of pulmonary metas- living tissue.27 By applying a nonionizing magnetic field, tasis in patients with a known or suspected bone a fraction of spinning hydrogen ions within a given tis- sarcoma.14,17,34 sue (primarily in water and lipid molecules) align them- selves along the line of magnetic field. When stimulated MRI with a radiofrequency (RF) pulse, these ions change The development of MRI over the past 2 decades alignment and tilt. The amount of tilt or flip is deter- has had tremendous impact on musculoskeletal mined by the strength and duration of the radiofre- imaging. The contrast resolution of MRI is nearly quency pulse. On returning to equilibrium, the ions 50 times that of conventional radiography and roughly generate radio waves that are detected by a receiver 10 times that of CT, while sparing the patient from coil. The strength of this signal is a function of the exposure to radiation.42 Additionally, the images can amount of hydrogen nuclei within the tissue. The fre- be manipulated by varying signal parameters that al- quency of the signal is proportional to the size of the low the characterization of tissue types. This key applied magnetic field. Protons in different chemical feature enables the anatomic, high-resolution eval- environments experience different magnetic fields uation of neoplasms and contributes tremendously because of interference from local electrons. The

442 Original Article

Cummings et al.

Images are obtained by applying a series of RF pulses. A receiving coil follows at a specified time, detecting the energy released from the realigning (relaxing) protons.31 Two key relaxation parameters are used: T1 and T2. T1 is the exponential time constant for a proton to return to equilibrium after receiving an RF pulse so that it can generate a full signal on subsequent pulses (the interval between pulses is termed the TR time). Rapid RF pulses gives weaker signals from the protons with a long T1 Figure 6 (A) Anteroposterior pelvic radiograph of a 13-year-old patient who complained of increasing left hip time.27 T1-weighted im- pain shows very subtle heterogenous lucency and increased density of the left ilium. (B) Technetium bone scan clearly shows increased activity in the entire left hemi-pelvis. Biopsy revealed Ewing’s sarcoma. ages are generated with a short TR time (400–600 frequency shift attributable to these fields is termed ms) and are best for evaluating the extent of a tumor chemical shift and helps explain the differing NMR sig- within bone marrow.43 T2 is the exponential time con- nal of hydrogen-1 in fat and water. On detection by the stant for a proton’s NMR signal to decay after it is gen- receiver coil, the signal is amplified and processed into erated. Protons with a short T2 can be differentiated the pixel grayscale level of the image.27,39 from those with a long T2 by delaying the detection A key advantage of MRI is the number of available of the NMR signal after the RF pulse. This detection mechanisms used to create contrast within an image. delay is referred to as the TE time.27 Images created us- Among the different pulse sequences used in generating a long TE time (>70 ms) are referred to as T2- ing images, spin echo sequences are used most often. weighted images and are most useful in determining cortical bone or soft-tissue involvement. Proton density images have long TR and short TE times. An important distinction between T1 and T2 images is the relative brightness of fluid on T2 images and darkness of fluid on T1 images (Figure 8).27 Most musculoskeletal tumors are very cellular and tend to behave like water. On T1-weighted images, tumors are often low to intermediate in signal intensity and can easily be identified against or within fat (including fatty marrow) but tend to blend with muscle. On T2-weighted images, many tumors are bright and easily identified.44 Additional pulse sequences are also used to generate images. Gradient echo imaging uses a series of smaller superimposed magnetic fields to generate the Figure 7 Computed tomography of the pelvis of a 73-year-old patient with chondrosarcoma shows erosion of the pubis and a soft-tissue mass NMR signal with shorter imaging times. Inversion with matrix calcification. recovery images use RF pulses applied in a particular

Original Article 443

Imaging of Bone Sarcomas

Figure 8 Coronal magnetic resonance images of the left proximal tibia of an 18-year-old patient with osteosarcoma shows the extent of the lesion within the bone. (A) On T1-weighted image, the tumor appears dark and the normal fatty marrow appears bright. (B) On fat-suppressed T2-weighted image, the tumor appears bright and the marrow appears dark.

order, effectively cancelling all signal normally pro- Nonetheless, MRI affords unparalleled tissue duced by fat.31 contrast, precise anatomic detail, and excellent sen- Exogenous sources of contrast can be applied to sitivity for imaging bone marrow, making it essen- alter the NMR signal. The most widely used contrast tial to the staging of known or suspected bone agents are the gadolinium (Gd) chelates, which are sarcomas. MRI is the best imaging modality for eval- 1000 times more visible than iodine x-ray contrast uating the extent of a lesion within bone and soft tis- agents. Gd metal asserts its effect by shortening the sues before and after chemotherapy or radiotherapy T1 relaxation of water. Hence, protons near a Gd (Figure 9).31,40,41,53–57 It is also the best test to evaluate atom will continue to yield a high signal on T1- the relationship between a tumor and adjacent weighted images.27 Gd-contrasted images may be useful in determining whether a lesion is solid or cystic and defining areas of viable tumor to help guide biopsy decisions. Despite the many advantages of MRI, several limitations remain, such as low specificity.45–52 The appearance of infection, benign tumors, malignant tumors, and trauma may be indistinguishable,26 and the appearance of structures that remain dark on all images (e.g., cortical bone, tendon, air, flowing blood) may be sim- ilar. Additionally, implanted metallic devices may be contraindicated for MRI, and care should be taken to document safety of such devices before proceeding. Finally, although MRI is highly accurate in defining the extent of a bone abnormality, CT and conventional ra- Figure 9 Axial MRI corresponding to the radiograph and bone scan of the diographs remain superior for visualizing bony detail patient in Figure 6 shows the extent of the tumor within soft-tissues and and matrix mineralization.26,27,31 the anatomic relationship of the tumor with the surrounding structures.

444 Original Article

Cummings et al.

Currently, the most commonly used isotope is 18F- labeled 2-fluoro-2-deoxyglucose (18FDG). Several features of 18FDG have been extremely useful. Intracellular transport does not distinguish it from glucose, uptake increases in cells with high metabolic activity (malignancies generally have higher rates of glycolysis than normal tissue), and low membrane per- meability results in intracellular accumulation. Other isotopes, such as fluorine-18, carbon-11, and 15-oxy- gen, have short half-lives precluding their use in clinical settings.27,39,66 Recently, dual-modality scanners capable of con- current CT or MRI and PET scanning have been de- veloped to allow integration of anatomic, functional, and metabolic information gathered from each imaging modality.27 This information is valuable because the detection of metabolically active tissue can assist with guided biopsy. In general, high-grade sarcomas and aggressive benign lesions show higher uptake than benign lesions.66,67 This trend becomes tenuous when Figure 10 CT-PET scan of a 23-year-old man shows tumor recurrence considering cartilaginous bone lesions, possibly be- in the soft-tissue posterior to the right scapula 3 years after chemotherapy and surgical resection of a right humeral Ewing’s sarcoma. The cause of the relatively low metabolic activity of carti- lesion (arrows) is difficult to see on the CT (A) but is readily apparent lage, which shows primarily anaerobic glycolysis.68–70 on the PET scan (B). Although highly sensitive, FDG PET unfortunately has relatively low specificity.71,72 High uptake has been anatomic structures.58–61 For complete evaluation of reported in patients with chronic inflammatory con- a suspected bone malignancy, standard spin echo ditions such as osteomyelitis or arthritis73,74 and in T1-weighted sequences, fast spin echo fat-suppressed many benign conditions such as acute fracture, fibrous T2-weighted or inversion recovery sequences, and dysplasia, and Paget’s disease.72,75–77 Therefore, the role post-contrast T1 fat-suppressed images are extremely of PET scanning in the initial evaluation and staging useful.44 The entire bone should be included to eval- of suspected bone malignancy remains investigational. uate for skip metastases, which is a poor prognostic The usefulness of PET scanning in monitoring for indicator.62–65 recurrent disease or response to treatment also has PET been varied in the literature. Some studies have PET scanning relies on the detection of 2 positively shown high sensitivities (93%–98%) for detecting charged photons (each 511 keV) that result from the recurrent or residual disease,78,79 whereas others have interaction of administered positrons (positively found both sensitivity and specificity of FDG PET in charged electrons) and native electrons inside living detecting recurrent tumors to be lower than those tissue. These 511-keV gamma rays (photons) are emit- for MRI.80 ted at 180° from each other and are registered by dual detectors encircling the body.66,67 By detecting these gamma rays at the same time, their spatial distribu- Summary tion can be reconstructed and corrected for tissue The appropriate application of imaging technology is attenuation signal loss. This allows quantification of of paramount importance for physicians evaluating the distribution of a radiopharmaceutical and its cor- and treating suspected bone malignancy. Each pa- relation with tissue metabolic activity. Uptake in PET tient’s workup begins with a thorough history and scanning is quantified using standardized uptake val- physical examination emphasizing the patient’s age, ues (SUVs). In general, an SUV of more than 3 sug- history of malignancy, and presence or absence of pain. gests an aggressive process. The lesion is first evaluated with plain radiographs,

Original Article 445

Imaging of Bone Sarcomas

which provide information about bone destruction, 10. Paulussen M, Ahrens S, Dunst J, et al. Localized Ewing tumor of tumor matrix, periosteal or endosteal reaction, lesion bone: final results of the Cooperative Ewing’s Sarcoma Study CESS 86. J Clin Oncol 2001;19:1818–1829. margins, associated soft-tissue changes, and lesion lo- 11. Rosen G, Caparros B, Huvos AG, et al. Preoperative chemotherapy cation. Clinical and radiographic findings should al- for osteogenic sarcoma: selection of postoperative adjuvant low formulation of a working differential diagnosis to chemotherapy based on the response of the primary tumor to pre- help determine the necessity for further imaging stud- operative chemotherapy. Cancer 1982;49:1221–1230. 12. Rosen G, Nirenberg A, Caparros B, et al. Osteogenic sarcoma: eight- ies. CT is often used to further evaluate matrix min- percent, three-year, disease-free survival with combination eralization and bony architecture for lesions in axial chemotherapy (T-7). Natl Cancer Inst Monogr 1981;56:213–220. locations and to evaluate pulmonary metastasis. 13. Rougraff BY, Simon MA, Kneisl JS, et al. Limb salvage compared Whole-body bone scanning can be used to evaluate le- with amputation for osteosarcoma of the distal end of the femur. A long-term oncological, functional, and quality-of-life study. J Bone sion activity and possible multifocal bone involve- Joint Surg Am 1994;76:649–656. ment. MRI is used to evaluate the extent of a lesion 14. Kerns LL, Simon MA. Surgical theory, staging, definitions and treat- within the bone and soft tissues and the relationship ment of musculoskeletal sarcomas. Surg Clin N Am 1983;63: of the lesion to surrounding anatomic structures. The 671–696. role of PET scanning remains undefined. It may prove 15. Pommersheim WJ, Chew FS. Imaging, diagnosis, and staging of bone tumors: a primer. Semin Roentgenol 2004;39:361–372. useful in staging (especially for detecting nonpul- 16. Unni KK. Dahlin’s Bone Tumors. General Aspects and Data on monary metastasis), evaluating response to 11,087 Cases, 5th ed. Philadelphia: Lippincott-Raven, 1996:7–8. chemotherapy, and detecting recurrent local or 17. Peabody TD, Gibbs CP, Simon MA. Current concepts review. metastatic disease during follow-up. Evaluation and staging of musculoskeletal neoplasms. J Bone Joint Surg Am 1998;80:1204–1218. 18. Mankin HJ, Lange TA, Spanier SS. The hazards of biopsy in patients with malignant primary bone and soft-tissue tumors. J Bone References Joint Surg Am 1982;64:1121–1127. 1. Bielack SS, Kempf-Bielack B, Delling G, et al. Prognostic factors in 19. Mankin HJ, Mankin CJ, Simon MA. The hazards of the biopsy, high-grade osteosarcoma of the extremities or trunk: an analysis of revisited. J Bone Joint Surg Am 1996;78:656–662. 1,702 patients treated on neoadjuvant Cooperative Osteosarcoma 20. Simon MA. Current concepts review. Biopsy of musculoskeletal Study Group protocols. J Clin Oncol 2002;20:776–790. tumors. J Bone Joint Surg Am 1982;64:1253–1257. 2. Bacci G, Ferrari S, Bertoni F, et al. Long-term outcome for patients 21. Simon MA. Diagnostic and staging strategy for musculoskeletal tu- with nonmetastatic osteosarcoma of the extremity treated at the mors. In: Evarts CM, ed. Surgery of the Musculoskeletal System, Istituto Orthpedico Rizzoli according to the Istituto Ortopedico 2nd ed. New York: Churchill Livingstone, 1990:4605–4634. Rizzoli/Osteosarcoma-2 protocol: an updated report. J Clin Oncol 2000;18:4016–4027. 22. Temple HT, Bashore CJ. Staging of bone neoplasms: an orthope- dic oncologist’s perspective. Semin Musculoskelet Radiol 2000;4: 3. Bacci G, Picci P, Ruggieri P, et al. The Istituto Rizzoli experience in 17–23. 127 patients treated preoperatively with intravenous methotrexate (high versus moderate doses) and intraarterial cisplatin. Cancer 23. Davies AM, Wellings RM. Imaging of bone tumors. Curr Opin Radiol 1990;65:2439–2553. 1992;4:32–38. 4. Goorin AM, Perez-Atayde A, Gebhardt M, et al. Weekly high-dose 24. Kricun ME. Parameters of diagnosis. In: Kricun ME, ed. Imaging of methotrexate and doxorubicin for osteosarcoma: the Dana-Farber Bone Tumors. Philadelphia: WB Saunders, 1993:2–45. Cancer Institute/the Children’s Hospital study III. J Clin Oncol 25. Lang P, Johnston JO, Arenal-Romero F, et al. Advances in MR im- 1987;5:1178–1184. aging of pediatric musculoskeletal neoplasms. Magn Reson Imaging 5. Kolb EA, Kushner BH, Garlick R, et al. Long-term event-free Clin N Am 1998;6:579–604. survival after intensive chemotherapy for Ewing’s family of tumors 26. Massengill AD, Seeger LL, Eckardt JJ. The role of plain radiography, in children and young adults. J Clin Oncol 2003;18:3423–3430. computed tomography, and magnetic resonance imaging in sarcoma 6. Link MP, Goorin MA, Miser AW, et al. The effect of adjuvant evaluation. Hematol Oncol Clin North Am 1995;9:571–604. chemotherapy on relapse-free survival in patients with osteosarcoma 27. Schnall M, Rosen M. Primer on imaging technologies for cancer. of the extremity. N Engl J Med 1986;314:1600–1606. J Clin Oncol 2006;24:3225–3233. 7. Mankin HJ, Hornicek FJ, Rosenberg AE, et al. Survival data for 28. Manaster BJ. Musculoskeletal oncologic imaging. Int J Radiat Oncol 648 patients with osteosarcoma treated at one institution. Clin Biol Phys 1991;21:1643–1651. Orthop Rel Res 2004;429:286–291. 29. Seeger LL, Widoff BE, Bassett LW, et al. Preoperative evaluation of 8. Marcus RB, Berrey BH, Graham-Pole J, et al. The treatment of osteosarcoma: value of gadopentetate dimeglumine-enhanced MR Ewing’s sarcoma of bone at the University of Florida: 1969 to 1998. imaging. AJR Am J Roentgenol 1991;157:347–351. Clin Orthop Rel Res 2002;397:290–297. 30. Wolf RE, Enneking WF. The staging and surgery of musculoskele- 9. Meyers PA, Heller G, Healey J, et al. Chemotherapy for non-metasta- tal neoplasms. Orthop Clin N Am 1996;27:473–481. tic osteogenic sarcoma: the Memorial Sloan-Kettering experience. 31. Sanders TG, Parsons TW III. Radiographic imaging of musculoskele- J Clin Oncol 1992;10:5–15. tal neoplasia. Cancer Control 2001;8:221–231.

446 Original Article

Cummings et al.

32. Enneking WF. A system of staging musculoskeletal neoplasms. Clin pressed T2-weighted, and STIR images. AJR Am J Roentgenol Orthop Rel Res 1986;204:9–24. 1994;162:215–221. 33. Olson PN, Everson LI, Griffiths HJ. Staging of musculoskeletal 53. Aisen AM, Martel W, Braunstein EM, et al. MRI and CT evalua- tumors. Radiol Clin N Am 1994;32:151–162. tion of primary bone and soft-tissue tumors. AJR Am J Roentgenol 34. Peh WC. The role of imaging in the staging of bone tumors. Crit Rev 1986;146:749–756. Oncol Hematol 1999;31:147–167. 54. Bloem JL, Taminiau AH, Eulderink F, et al. Radiologic staging of 35. Teo HE, Peh WC. The role of imaging in the staging and treatment primary bone sarcoma: MR imaging, scintigraphy, angiography, planning of primary malignant bone tumors in children. Eur Radiol and CT correlated with pathologic examination. Radiology 1988; 2004;14:465–475. 169:805–810. 36. Brien EW, Mirra JM, Kerr R. Benign and malignant cartilage tu- 55. Bohndorf K, Reiser M, Lochner B, et al. Magnetic resonance imag- mors of bone and joint: their anatomic and theoretical basis with an ing of primary tumors and tumor-like lesions of bone. Skeletal Radiol emphasis on radiology, pathology and clinical biology. I. The in- 1986;15:511–517. tramedullary cartilage tumors. Skeletal Radol 1997;26:325–353. 56. Boyko OB, Cory DA, Cohen MD, et al. MR imaging of osteogenic and Ewing’s sarcoma. AJR Am J Roentgenol 1987;148:317–322. 37. Simon MA, Krichner PT. Scintigraphic evaluation of primary bone tumors. Comparison of technetium-99m phosphate and gallium 57. Cohen IJ, Hadar H, Schreiber T, et al. Primary bone tumor citrate imaging. J Bone Joint Surg 1980;62:759–764. respectability: the value of serial MRI studies in the determination of the feasibility, timing, and extent of tumor resection. J Pediatr 38. Hounsfield GN. Computerized transverse axial scanning (tomogra- Orthop 1994;14:781–787. phy): 1. Description of system. Br J Radiol 1973;46:1016–1022. 58. Petasnick JP, Turner DA, Charters JR, et al. Soft-tissue masses of 39. Barentsz J, Takahashi S, Oyen W, et al. Commonly used imaging the locomotor system: comparison of MR imaging with CT. Radiology techniques for diagnosis and staging. J Clin Oncol 2006;24: 1986;160:125–133. 3234–3244. 59. Pettersson H, Gillespy T, Hamlin DJ, et al. Primary musculoskele- 40. Sundaram M, McDonald DJ. Magnetic resonance imaging in the tal tumors: examination with MR imaging compared to conven- evaluation of the solitary bone tumor. Curr Opin Radiol 1990; tional modalities. Radiology 1987;164:237–241. 2:697–702. 60. Vanel D, Verstraete KL, Shapeero LG. Primary tumors of the mus- 41. Sundaram M, McGuire MH. Computed tomography or magnetic culoskeletal system. Radiol Clin N Am 1997;35:213–237. resonance for evaluating the solitary tumor or tumor-like lesion of 61. Zimmer WD, Berquist TH, McLeod RA, et al. Bone tumors: mag- bone? Skeletal Radiol 1988;17:393–401. netic resonance imaging versus computed tomography. Radiology 42. Richardson ML, Gillespy T. Magnetic resonance imaging. In: Kricun 1985;155:709–718. ME, ed. Imaging of Bone Tumors. Philadelphia: SB Saunders, 62. Enneking WF, Kagan A. The implications of “skip” metastases in 1993:358–446. osteosarcoma. Clin Orthop Rel Res 1975;111:33–41. 43. Berquist TH. Magnetic resonance imaging of musculoskeletal 63. Malawer MM, Dunham WK. Skip metastases in osteosarcoma: re- neoplasms. Clin Orthop Rel Res 1989;244:101–118. cent experience. J Surg Oncol 1983;22:236–245. 44. Jaramillo D, Laor T, Gebhardt MC. Pediatric musculoskeletal neo- 64. Sajadi KR, Heck HK, Neel MD, et al. The incidence and prognosis plasms. Evaluation with MR imaging. Magn Reson Imaging Clin N of osteosarcoma skip metastases. Clin Orthop Rel Res 2004;426: Am 1996;4:749–770. 92–96. 45. Assoun J, Richardi G, Railhac JJ, et al. Osteoid osteoma: MR imag- 65. Wuisman P, Enneking WF. Prognosis for patients who have osteosar- ing versus CT. Radiology 1994;18:981–984. coma with skip metastasis. J Bone Joint Surg Am 1990;72:60–68. 46. Benedikt RA, Jelinek JS, Kransdorf MJ, et al. MR imaging of soft- 66. Aoki J, Endo K, Watanabe H, et al. FDG-PET for evaluating mus- tissue masses: role of gadopentetate dimeglumine. J Magn Reson culoskeletal tumors: a review. J Orthop Sci 2003;8:435–441. Imaging 1994;4:485–490. 67. Feldman F, van Heertum R, Manos C. 18FDG PET scanning of 47. Constable RT, Smith RC, Gore JC. Signal-to-noise and contrast in benign and malignant musculoskeletal lesions. Skeletal Radiol fast spine echo (FSE) and inversion recovery FSE imaging. J Comput 2003;32:201–208. Assist Tomogr 1992;16:41–47. 68. Gilkey FW, Moser RP. The biology of cartilage. In: Moser RP, ed. 48. Holscher HC, Bloem JL, van der Woude HJ, et al. Can MRI pre- Cartilaginous Tumors of the Skeleton. Philadelphia: Hanley and dict the histopathological response in patients with osteosarcoma Belfus, 1990:1–7. after the first cycle of chemotherapy? Clin Radiol 1995;50: 69. Lee FY, Yu J, Chang SS, et al. Diagnostic value and limitations of 384–390. fluorine-18 fluorodeoxyglucose positron emission tomography for 49. Holscher HC, van der Woude HJ, Hermans J, et al. Magnetic reso- cartilaginous tumors of bone. J Bone Joint Surg Am 2004;86: nance relaxation times of normal tissue in the course of chemother- 2677–2685. apy: a study in patients with bone sarcoma. Skeletal Radiol 70. Poole AR. Physiology of cartilage formation, function and destruc- 1994;23:181–185. tion. In: Cruess RC, ed. The Musculoskeletal System. Embryology, 50. Li KC, Hopkins KL, Moore SG, et al. Magnetization transfer con- Biochemistry and Physiology. New York: Churchill Livingstone, trast MRI of musculoskeletal neoplasms. Skeletal Radiol 1995;24: 1982:286–230. 21–25. 71. Dehdashti F, Siegel BA, Griffeth LK, et al. Benign versus malignant 51. Ma LD, Frassica FJ, Scott WW, et al. Differentiation of benign and intraosseous lesions: discrimination by means of PET with malignant musculoskeletal tumors: potential pitfalls with MR imag- 2-[F-18]fluoro-2-deoxy-D-glucose. Radiology 1996;200:243–247. ing. Radiographics 1995;15:349–366. 72. Schulte M, Brecht-Krauss D, Heymer B, et al. Grading of tumors 52. Mirowitz SA, Apicella P, Reinus WR, et al. MR imaging of bone and tumorlike lesions of bone: evalution by FDG PET. J Nucl Med marrow lesions: relative conspicuousness on T1-weighted, fat-sup- 2000;41:1695–1701.

Original Article 447

Imaging of Bone Sarcomas

73. Guhlmann A, Brecht-Krauss D, Suger G, et al. Chronic osteomyelitis: 77. Strauss LG. Fluorine-18 deoxyglucose and false-positive results: detection with FDG PET and correlation with histopathologic find- a major problem in the diagnostics of oncological patients. Eur J ings. Radiology 1998;206:749–754. Nucl Med 1996;23:1409–1415. 74. Palmer WE, Rosenthal DI, Schoenberg OI, et al. Quantification of 78. Garcia JR, Kim EE, Wong FCL, et al. Comparison of fluorine-18-FDG inflammation in the wrist with gadolinium-enhanced MR imaging PET and technetium-99m-MIBI SPECT in evaluation of muscu- and PET with 2-[F-18]-fluoro-2-deoxy-D-glucose. Radiology loskeletal sarcomas. J Nucl Med 1996;37:1476–1479. 1995;196:647–655. 79. Kole AC, Nieweg OE, van Ginkel RJ, et al. Detection of local re- 75. Akoi J, Sone S, Fujioka F, et al. MR of enchondroma and chon- currence of soft-tissue sarcoma with positron emission tomogra- drosarcoma: rings and arcs of Gd-DTPA enhancement. J Comput phy using [18F]fluorodeoxyglucose. Ann Surg Oncol 1997;4: Assist Tomogr 1991;15:1011–1016. 57–63. 76. Kern KA, Brunetti A, Norton JA, et al. Metabolic imaging of 80. Lucas JD, O’Doherty MJ, Wong JCH, et al. Evaluation of fluo- human extremity musculoskeletal tumors by PET. J Nucl Med rodeoxyglucose positron emission tomography in the management 1988;29:181–186. of soft-tissue sarcomas. J Bone Joint Surg Br 1998;80:441–447.