Imaging of Paget's Disease and Fibrous Dysplasia of Bone

Abstract

Paget's disease of bone is a frequently encountered incidental finding on radiographs and scintigrams. Radiographically, the intermediate phase of Paget's disease expressed by sclerosis, lysis, thickened trabecula, and cortices is not usually a diagnostic challenge when encountered in the pelvis, calvaria, ends of long tubular bones, and vertebra. The purely lytic form of Paget's disease, the sclerotic form without bone expansion, and the combination of lytic Paget's in an a-typical location such as the tibial tubercle continue to be diagnostically confounding in comparison with its occurrences in characteristic locations in the intermediate phase of the disease. This review will be confined to the atypical form of Paget's disease and the role of MRI in separating uncomplicated Paget's disease from its well-recognized complications, albeit rare, of superimposed sarcoma. Virtually any bone can be affected by Paget's disease, and because the disease is more often asymptomatic than symptomatic, radiologists, perhaps more than most physicians, frequently encounter the disease in well-recognized locations such as the pelvis or vertebra and in less well-identified or -recognized locations such as the sacral ala or metacarpals. This is usually the case when imaged for staging of a malignancy or symptoms after trauma. Paget's disease predominates in the pelvis, lumbar segment of the vertebral column, and long bones, with a preference for its ends. The disease has three distinct phases, lytic intermediate, and sclerotic, and a fourth phase when exacerbation of the lytic phase occurs after the intermediate and sclerotic phases have been established. The purely lytic phase of the disease is considered its rarest clinical manifestation.1 Because of its aggressive appearance and afflicted individuals often being in the fifth decade of life or beyond at the time of lesion discovery, this form of Paget's often engenders the differential diagnosis of metastasis or lymphoma. Technetium scintigraphy shows marked increases of radio tracer accumulation, indistinguishable from a malignant disease process. MR imaging, however, may, in one of those rare clinical situations, permit more specific characterization than the radiograph staving, off an unnecessary biopsy. In osteolytic Paget's uncomplicated by either fracture or sarcoma, the fatty marrow signal is preserved on the T1-weighted MRI sequences.2 This is one of the very few situations where an osteolytic “lesion” on a radiograph does not show the usually encountered low signal on T1-weighted MRI sequences. Because the destruction of bone in lytic Paget's is caused by resorption and not infiltration, fatty marrow and its corresponding MRI signal are preserved. A clue to the radiographic diagnosis of lytic Paget's disease of a long bone is the “blade of grass” appearance, which in association with MRI, should prevent an unnecessary biopsy. The role of positron emission tomography (PET) in asymptomatic osteolytic Paget's is uncertain, although it has been shown that, in the usually encountered asymptomatic Paget's disease, there is no uptake on PET, and in the lesions that did show a mild uptake, the PET abnormality correlated with elevated serum alkaline phosphatase.3 This phase of the disease is usually not a diagnostic challenge unless complicated by superimposed osteolysis, which may be cortical or cancellous in location and when a sarcoma may be mimicked. For the reasons stated above, MRI permits separation of a sarcoma from superimposed osteolysis.4 The sclerotic phase of Paget's may occur without bone expansion creating a diagnostic dilemma in patients with known prostatic or breast carcinoma. This is especially so when the abnormality occurs in sites favored by metastases such as vertebra and pelvis. The signal features of MRI previously referred to may not be helpful in this situation because the MR signal returned by sclerotic lesions of bone, irrespective of the cause for the sclerosis, would be low in all MRI sequences. MRI may, however, serve to exclude a soft tissue mass. Technetium scintigraphy is also likely to be of limited value because it is likely to show increased radio-tracer accumulation. Although not the basis of a formal study anecdotal experience suggests that PET scanning may help, in that sclerotic Paget's is unlikely to be hypermetabolic—a feature associated with osseous metastases. Cortical insufficiency fractures usually seen in the femur and tibia are more common than complete fractures. These are seen as horizontal lines that may be single or multiple, favoring the convex aspects of bone unlike osteomalacic insufficiency fractures, which favor the concave aspects of long bones. The subtrochanteric femur is a common location. The prevailing view is that ∼1% of patients with Paget's disease may be afflicted with a superimposed sarcoma.5 MRI usually serves to separate Paget's sarcoma from nonmalignant Paget's disease.4 Although not formally studied, our understanding of MRI “neoplastic” signal characteristics would suggest that MRI may not help to distinguish Paget's sarcoma from giant cell tumor in Paget's or other neoplasm co-existing with Paget's disease. Pseudosarcoma in Paget's disease is a radiographic appearance that strongly suggests sarcoma because of a soft tissue mass and quite profound periosteal proliferation.6, 7 The morphological appearance has been likened to “poultice.” In two of three such cases reported with MRI, the signal characteristics of the lesion on MRI suggested that the lesions were not sarcomatous.6, 7 Rapid osteolysis of bone in a patient with Paget's disease may be observed in patients who are immobilized usually because of a fracture. The osteolysis is often so dramatic and extensive that an acute osteomyelitis or malignant disease process is considered. Often the limb in question may be casted, precluding a clinical examination. We are not aware of MRI being performed in this setting but do believe that, because this is a form of acute osteoporosis, the fatty marrow signal would largely be preserved, especially away from the fracture where osteolysis may be quite profound, precluding both osteomyelitis and sarcoma as causes for the massive osteolysis. Fibrous dysplasia is a benign, medullary, fibro-osseous lesion that may involve one (monostotic) or more bones (polyostotic) and is encountered worldwide, in children and adults, in all racial groups, and with an equal sex distribution. The monostotic form is about six times as common as the polyostotic form. Preponderance of females being afflicted by fibrous dysplasia is usually confined to McCune-Albright syndrome (MAS),8 which is a sporadically occurring disorder consisting of polyostotic fibrous dysplasia, café-au-lait spots, and hyperfunctioning endocrinopathies encountered in ∼3% of patients with fibrous dysplasia. Mazabraud's syndrome is the association of intramuscular myxoma and fibrous dysplasia.9 The myxomas may be encountered in the monostotic form10 but has more frequently been reported in the polyostotic form of the disease. The World Health Organization (WHO) classifies fibrous dysplasia as a tumor of undefined neoplastic nature based on reports of clonal chromosomal aberrations.11 Activating mutations in the GNAS1 gene have been shown in the monostotic and polyostotic forms of the disease.12 The most common sites of involvement are rib, femur, tibia, and humerus. Fibrous dysplasia is considered the most common benign neoplasm of the rib. Craniofacial involvement is common, affecting one half of the patients with polyostotic disease and 10% of those with monostotic disease. The polyostotic form of fibrous dysplasia may present with a monomelic distribution, affecting one side of an entire extremity, usually lower. This pattern of distribution is a feature that it shares with Ollier's disease. Vertebral and hand involvement is exceptional. Fibrous dysplasia is primarily a radiographic diagnosis that rarely needs a confirmatory imaging study. Like Paget's disease, the finding is often incidental, and the diagnosis is presumptive. The radiographic features are varied and include a variety of patterns that may be lucent, sclerotic, mixed, or resembling ground glass. In weight-bearing bones, abnormal modeling may be encountered, and in its most extreme form, a “shepherd crook's” deformity may be seen in the femur. Abnormal bone modeling and complete fractures are a more common feature of the polyostotic form of the disease. Bone expansion may be seen in rib and craniofacial bones. Occasionally, bone expansion may be caused by the development of secondary aneurysmal bone cysts, a not uncommon finding since the advent of MRI. Calcification in fibrous dysplasia is an occasional finding. When this calcification is profound or extensive, it has led to the confusing pathological term of fibro-cartilaginous dysplasia. This is a redundant term.13 Calcification in fibrous dysplasia should be considered a part of the spectrum of disease as is secondary aneurysmal bone cysts or lesional fat. The biological behavior of fibrous dysplasia with calcification is similar to those lesions without calcification. An entity, liposclerosing myxofibrous tumor (LSMFT), with imaging features similar to fibrous dysplasia, has been described by authors, suggesting that the entity may be premalignant.14 Most authorities consider LSMFT to be fibrous dysplasia in a particular phase of its evolution, and that it may be a premalignant condition has little supportive evidence. The true incidence of malignancy de novo in fibrous dysplasia (i.e., no previous radiation) is difficult to estimate because of the innumerable cases of presumptively diagnosed fibrous dysplasia (on imaging) that do not come to biopsy. An entity that is malignant but likely to be misdiagnosed as fibrous dysplasia histologically is low-grade central osteosarcoma. One imaging feature that helps distinguish fibrous dysplasia from low-grade central osteosarcoma is that, in fibrous dysplasia of long bones or pelvis, cortical effacement/destruction with or without an adjacent soft tissue mass should not be seen. If the overall radiographic pattern is that of fibrous dysplasia but is associated with cortical destruction (seen radiographically or on MRI), the diagnosis of low-grade central osteosarcoma should be considered.15 Osteofibrous dysplasia is an abnormality predominantly of the cortices with some cancellous bone involvement almost always seen only in the tibia and fibula. Some reports have suggested that this entity may be related to adamantinoma, a low-grade malignancy, because of some shared histological features, and that some cases of osteofibrous dysplasia may progress to adamantinoma. This is a controversial subject and remains under review.16 For practical purposes, osteofibrous dysplasia is seen in the first two decades of life, whereas adamantinoma is seen in the third decade or after; deformity with bowing of the tibia is a feature of osteofibrous dysplasia and not usually seen with adamantinoma. Osteofibrous dysplasia does not metastasize, whereas adamantinoma may rarely do so. Surgery for osteofibrous dysplasia is usually for cosmetic reasons, whereas biopsy and excision is mandatory for adamantinoma. The MRI features of fibrous dysplasia reflect the heterogeneity of its histological composition, which in variable amounts comprise fibrovascular tissue, osteoid, cartilage, fat, and secondary aneurysmal bone cysts. MRI may allow for suggesting the diagnosis of low-grade central osteosarcoma when the radiographic features suggest fibrous dysplasia, if cortical effacement with or without a soft tissue mass is identified. Soft tissue myxoma seen in Mazabraud's syndrome has characteristic signal features and a morphology of a conventional myxoma, and the diagnosis of Mazabraud's syndrome is suggested when the examiner discovers a soft tissue myxoma in a patient with known fibrous dysplasia. To our knowledge, development of a myxosarcoma in a patient with Mazabraud's syndrome has not been described.