The radiograph remains the initial imaging study for evaluating both bone and soft tissue lesions. For osseous lesions, it is almost invariably the most diagnostic. The radiograph accurately predicts the biologic activity of a bone lesion, which is reflected in the appearance of the margin of the lesion, and the type and extent of accompanying periosteal reaction. In addition, the pattern of associated matrix mineralization may be a key to the underlying histology (e.g., cartilage, bone, fibro-osseous) [2–5]. Although other imaging modalities (MR and CT) are superior to radio — graphs in staging a bone lesion, the radiograph remains the best modality for establishing a diagnosis, for formulating a differential, and for accurately assessing the biologic activity (separating benign from malignant lesions). In many cases, as in patients with fibroxanthoma (nonossifying fibroma), fibrous dysplasia, osteochondroma or enchondroma, radiographs may be virtually pathognomonic, and no further diagnostic imaging is required. In other cases, despite not having an unequivocal diagnosis, a benign-appearing, asymptomatic lesion may require only continued radiographic follow-up in order to document long-term stability.

Radiographs are typically considered unrewarding in the assessment of soft tissue masses; however, a recent study of 281 patients showed calcification in 27%, bone involvement in 22% and fat in 11% of cases [6]. Such features may be essential in establishing an appropriate imaging diagnosis, establishing a differential, and in assessing malignant potential. Radiographs may also be diagnostic of a palpable lesion caused by an underlying skeletal deformity (such as exuberant callus related to prior trauma) or exostosis, which may masquerade as a soft tissue mass. The soft tissue calcifications and/or ossification identified on radiographs can be suggestive, and at times highly characteristic, of a specific diagnosis. For example, they may reveal the phleboliths within a hemangioma, the juxta-articular osteocartilaginous masses of synovial chondromatosis, the peripherally more mature ossification of myositis ossificans, or the characteristic bone changes of other processes with associated soft tissue involvement. When not characteristic of a specific process, soft tissue calcification can suggest certain diagnoses. For example, nonspecific dystrophic calcifications within a slowly growing lower extremity mass in an adolescent or young adult should suggest a synovial sarcoma as the diagnosis of exclusion.

MR imaging is the preferred modality for evaluating soft tissue lesions. It provides superior soft tissue contrast, allows multiplanar image acquisition, obviates the need for iodinated contrast agents or for ionizing radiation, and is devoid of streak artifact commonly encountered with CT imaging [7–10]. When clinical findings are equivocal, MR imaging evaluation can confirm the presence of a soft tissue lesion or reassuringly identify a suspected „bump” or „mass” as normal tissue [11]. Whereas MR imaging has emerged as the preferred advanced imaging modality for the evaluation of soft tissue lesions, CT and MR imaging are often complimentary modalities for the evaluation of primary osseous tumors [10]. In a study by Tehranzadeh et al., the information obtained from CT and MR imaging was additive in 76% of malignant primary bone neoplasms [10]. CT scanning is superior to MR imaging for the detection and characterization (osteoid or chondroid) of matrix mineralization, cortical involvement and periosteal reaction [9, 10]. Although some studies suggest that CT and MR are equivalent for the evaluation of the local extent of tumor [12], most studies note the superiority of MR imaging [9, 10, 13–15].

Contrast administration may be a useful adjunct in the assessment of both bone and soft tissue lesions; however, its use is usually dictated by the objectives of the examination. For example, in the CT evaluation of osseous lesions, it is typically not required to assess the presence or character of matrix mineralization, periosteal reaction or marginal sclerosis. On MR imaging, many lesions are well assessed without contrast. As a rule, we find it of little value in the assessment of lipoma or the usual atypical lipomatous tumor (atypical lipoma). In some cases, contrast administration can also cause confusion, blurring the distinction of tumor from peritumoral edema and normal from abnormal marrow [14, 32]. Nevertheless, it may provide essential information and we find contrast imaging especially important in the assessment of lesions containing hemorrhage, myxomatous areas, or necrotic or cystic regions. Only vascularized tissue enhances; therefore, contrast enhanced imaging may be quite useful in directing biopsy to the solid, enhancing portions of a lesion, the portion of the lesion that harbors the diagnostic tissue, as opposed to the cystic, necrotic or hemorrhagic nondiagnostic components. Additionally, the pattern of enhancement may also provide clues to the diagnosis, as in the characteristic peripheral and septal enhancement pattern seen with hyaline cartilage neoplasms.

Contrast-enhanced imaging is not without a price. The use of intravenous contrast increases the length and cost of the examination. Although contrast-enhanced MR imaging may provide additional information, it usually does not increase lesion conspicuity nor replace the diagnostic value of T2-weighted imaging [33]. While MR contrast agents are safer than those used with CT, there is a small, but real, incidence of untoward reactions including hypotension, laryngospasm, bronchospasm and anaphylactic shock [34–36]. Most recently, the association of nephrogenic systemic fibrosis (NSF) with the use of gadolinium-based contrast agents has focused greater attention on its routine administration [37–39]. NSF was first reported in 2000 as a scleroderma-like fibrotic skin disorder in patients with renal insufficiency [38–40]. In a large 10-year study of 8,997 patients receiving gadolinium- based contrast, 15 (0.17%) developed NSF, all of whom had renal failure with glomerular filtration rates of less than 30 mL/min [37]. In a study at four US university tertiary care centers involving over 216,000 patients, NSF was found to be more than 15 times more common with gadodiamide (Omniscan™) than gadopentetate dimeglumine (Magnevist®) [38]. The development of NSF has also been associated with high cumulative doses of gadolinium-based contrast, leading to the recommendation that half-strength contrast be given to patients with glomerular filtration rates of less than 60 mL/min [37, 41–43]. Consequently, contrast-enhanced imaging should be reserved for those cases in which the results influence patient management.

Bone scintigrapy with 99mTc phosphate compounds remains the modality of choice for the identification of multifocal osseous disease. Positron emission tomography (PET) has established itself as the gold standard for metabolic imaging, and while it has not replaced skeletal scintigraphy, it has several intrinsic advantages. It has inherently superior spatial resolution and routinely includes tomographic images with multiplanar capability. Additionally, unlike skeletal scintigraphy, which primarily evaluates only the skeletal system, PET detects disease in both the osseous structures as well in the soft tissues, allowing detection of not only the primary tumor, but pulmonary, as well as nodal, metastasis [44]. Moreover, PET detects the presence of tumor directly, by reflecting its metabolic activity, rather than indirectly as with conventional scintigraphy, by demonstrating tumor involvement due to its increased bone mineral turnover [44].

The staging of musculoskeletal tumors is one of the primary functions of imaging. Accurate staging is essential for appropriate planning of therapy and establishing of prognosis. Simply stated, the purpose of a staging system is to provide a standard manner in which to readily communicate the state of a malignancy. Accurate staging is essential to (a) incorporate the most significant prognostic factors into a system that describes progressive degrees of risk to which a patient is subject, (b) delineate progressive stages of disease that have specific implications for surgical management, and (c) provide guidelines to the use of adjunctive therapies [45]. The staging systems most commonly used are the Enneking system and the staging system of the American Joint Committee. Staging requires a knowledge of the histologic grade of the lesion, as well as its anatomic extent, and can only be completed following biopsy.

As previously noted, the radiograph remains the most diagnostic imaging study. While more advanced imaging provides the information needed for accurate staging, diagnosis (or differential diagnosis) begins with the radio — graph and is based on the morphology of the lesion, the location of the lesion and the age of the patient. Morphology characterizes the presence and character of the lesion’s margin and periosteal reaction; features which are a function of the interaction between the lesion and the host, and are a key to biologic behavior. Morphology also includes an analysis of matrix. If present, a mineralized matrix may be a key to the lesion histology

The location of a lesion is a major consideration in establishing a diagnosis and structuring a differential diagnosis. There are two components to the location of an osseous lesion: the anatomic location (which bone) and the lesion’s location within the bone (epiphysis, meta — physis, metadiaphysis or diaphysis). The latter is critical in assessment of long bone lesions and is an essential principle of diagnosis pictorially presented in Fig. 1. Age is also a critical consideration, since specific lesions tend to occur in specific age groups (Tables 3, 4) [47]. For example, Langerhans cell histiocytosis localized to bone typically occurs in children and adolescents, while osseous lymphoma is most often encountered in mature adults, usually in the sixth and seventh decades. The final diagnosis, or differential, is then based on the integration of all of the above factors, taken in consideration of the prevalence of tumors within the population.

Despite the superiority of MR imaging in delineating soft-tissue tumors, it remains limited in its ability to precisely characterize them, with most lesions demonstrating a nonspecific appearance [48, 49]. There are instances, however, in which a specific diagnosis may be made or strongly suspected on the basis of MR imaging features (Box 3). This is usually done on the basis of lesion signal intensity, pattern of growth, location and associated „signs” and findings. The MR imaging appearance of these lesions has been well reported, and is not reviewed here.

DeSchepper et al. [50] performed a multivariate statistical analysis of ten imaging parameters, individually and in combination. These researchers found that malignancy was predicted with the highest sensitivity when lesions had high signal intensity on T2-weighted images, were larger than 33 mm in diameter, and had heterogeneous signal intensity on T1-weighted images. The signs that had the greatest specificity for malignancy included tumor necrosis, bone or neurovascular involvement, and mean diameter of more than 66 mm. In a recent study of 548 patients by Gielen and colleagues [51], in which imaging and clinical data were available, an accuracy of 85% was reported in differentiating between benign and malignant lesions.

When a specific diagnosis is not possible, it is often useful to formulate a suitably ordered differential diagnosis on the basis of imaging features, suspected biological potential and a knowledge of tumor prevalence based on the patient age and lesion anatomic location. This can be further refined by considering clinical history and radiologic features, such as pattern of growth, signal intensity and localization (subcutaneous, intramuscular, intermuscular, etc.). The most common malignant and benign lesions, by tumor location and patient age, are shown in Tables 5 and 6.