Liposclerosing Myxofibrous Tumor: A Review

Liposclerosing myxofibrous tumor typically presents in the fourth decade of life, with an equal sex distribution.^2,3,7  The reported age range is broad, ranging from 15 to 80 years.^3,5  Lesions are typically found incidentally on radiographic imaging, but patients can present with bone pain or pathologic fracture.^2,3  The tumor shows a strong predilection for the proximal femur, with 80% to 90% occurring in the intertrochanteric region.³  Other nonhip sites of presentation include the tibia, humerus, and rib.^2,3  The initial clue to diagnosis of LSMFT is a distinct set of radiographic features. This lesion is characteristically a well-defined geographic lytic lesion with a sclerotic border, suggesting an indolent growth pattern (Figure 1).³  The lesion can show expansive remodeling and mixed patterns of mineralization. Magnetic resonance images display a mildly heterogeneous lesion, with greater signal intensity than fat.³  Given this low-grade radiographic appearance, most specimens from this tumor arrive as bone fragments after curettage; hence, the gross picture of this tumor must be derived from the radiographic features.

Microscopically, LSMFT displays a variety of histologic patterns, leading many authors to question its origin as a primary lesion in favor of a traumatic or ischemic variant of another well-defined lesion or lesions. Various patterns of this lesion include areas of myxofibrous tissue with low cellularity, ischemic-pattern ossification, curvilinear trabeculae of woven bone (somewhat FD-like), woven bone with irregular mosaic lines of calcification (pseudo-Paget bone; Figure 2, A and B), lipomatous areas, and occasionally areas of cartilage.^3,4  Additional features include areas of fibrous tissue ossification and areas of fibrous tissue admixed with foamy macrophages (Figure 2, C through F). Significant cytologic atypia is not present.

The origin of this unique tumor is unclear and remains a source of discussion. Ragsdale²  suggests that this tumor arises in childhood, undergoing many subsequent changes throughout life. This origin, in his view, leads to the variety of histologic patterns seen in this tumor. This analysis stems from the fact that the tumors in teens and young adults have a predominantly fibroxanthomatous component with none of the other characteristic patterns.

Kransdorf et al³  analyzed the radiographic features of 39 tumors with histologic criteria suggestive of LSMFT. They suggest that this tumor arises as a result of involutional and ischemic changes to a lipogenic lesion of bone with superimposed bone proliferative changes. It is in this ischemic and reactive proliferative background that malignant transformation is thought to arise. Although Kransdorf et al³  accept that the various histologic patterns in this tumor could all be seen in FD, they suggest that the histologic pattern of LSMFT goes far beyond the standard description of FD. This would make routine recognition as a manifestation of FD difficult for most diagnostic pathologists. Furthermore, they argue that the discovery of an increased incidence of malignant transformation (4 cases in their series) compared with FD also supports the need for classification of this tumor as a distinct entity to encourage close monitoring and follow-up.

Heim-Hall and Williams⁴  propose another hypothesis, that these lesions are simply traumatized variants of FD. They suggest that the wide variety of patterns seen in this tumor represents the end result of repeated reaction to stresses on a bone altered by a fibro-osseous proliferation such as FD. The location distribution of this lesion lends some support to this theory, given that the weight-bearing stresses are considerably higher in the proximal femur than at other sites. Further substantiating this is a study by Matsuba et al,⁶  who found activating Gs-α mutations at the Arg²⁰¹ codon—mutations found in cases of monostotic and polyostotic FD—in several of these tumors. Dattilo and McCarthy⁵  have reviewed their experience with 33 cases of putative LSMFT and conclude on the basis of histologic evidence in 18 that the overall features fit with an underlying diagnosis of FD (12 cases) or intraosseous lipoma (6 cases) and that therefore the term should be abandoned, because use of the term could lead to unwarranted treatment. Further complicating this discussion are 2 well-studied cases reported by Corsi et al⁸  in which on the one hand an FD-like lesion with some histologic overlap with features suggesting LSMFT in the proximal femur did not have any identifiable Gs-α mutations after exhaustive study, and also could be shown to have arisen in adulthood, whereas a second lesion with more characteristic LSMFT morphology in an adult dancer did have an identifiable mutation. Also, credible evidence that LSMFT can occur as a de novo lesion in adulthood challenges the contention that the lesion simply represents monostotic FD occurring in the femur.

So, should use of the term LSMFT be abandoned? This is in some ways reminiscent of earlier discussions as to whether aneurysmal bone cyst exists as a primary lesion or is simply a reaction pattern observed in several underlying lesions that are at times obscured or obliterated by the metamorphosis. The real question is whether the distinction remains useful while further evidence accumulates as to the process(es) of its pathogenesis. Datillo and McCarthy⁵  argue that patients may be harmed by overzealous biopsies or curettage in settings where none would be warranted without the LSMFT specter of increased malignancy looming over them. But other studies purport directly observed transformation from LSMFT morphology to malignancy.⁷  This issue forms the principal support of its clinical utility—the apparent difference in malignant potential in follow-up of LSMFT as opposed to either FD or intraosseous lipoma. We are not at this point fully satisfied that all the purported malignant transformations represent instances of initial misdiagnosis of LSMFT or misdiagnosis of malignancy⁹  and that the mixed morphologic milieu, whether a primary lesion or a reaction pattern, may represent fertile ground for subsequent more serious alterations.⁷  It may be, as DNA sequencing drops in cost, that larger series of cases beyond the 2 of Matsuba et al⁶  and the 2 of Corsi et al⁸  with such testing will accrue and clarify the molecular landscape, allowing us to separate out the cases of FD with GNAS mutations from LSMFT, perhaps along with other potential mimics, and define whether this designation can then safely be abandoned. For the time being, we continue to think that the designation serves a purpose, and is helpful for pathologists to understand.

So whether it exists as a distinct pathologic entity, a unique variant of a more well-known lesion, or a reaction pattern imposed on a variety of precursor lesions in a fairly specific location, this lesion poses a distinct diagnostic challenge morphologically, necessitating the need for consideration in the differential diagnosis of fibro-osseous lesions.

Because of its location and various histologic patterns as has been described, LSMFT can mimic a variety of lesions. The principal differential diagnosis for this tumor includes FD, intraosseous lipoma, and nonossifying fibroma. Fibrous dysplasia, first described by Lichtenstein¹⁰  in 1938, is a noninherited developmental anomaly of bone in which normal bone marrow is replaced by a benign proliferation of fibro-osseous tissue. By definition, the lesions first appear in childhood, but may not manifest clinically until adulthood. The ossification is in a woven bone pattern, lacking any osteoblastic rimming, and often characterized by wandering curved, Chinese character–like bony trabeculae. This entity can occur in a single bone (monostotic) or multiple bones (polyostotic). When accompanied by endocrinopathies and skin pigmentation, the polyostotic form is known as McCune-Albright syndrome.¹¹  Patients typically are identified incidentally by radiographic exam for other reasons, but can also present with pain or pathologic fracture.

Although early authors²  suggested that FD is rare in the femur, other literature directly contradicts this; in fact, this lesion occurs quite commonly in the proximal femur.¹¹  This lesion presents radiographically as an intramedullary, well-defined expansile lesion that can be radiolucent or sclerotic. It typically exhibits a ground-glass appearance due to the altered bone spicules within the tumor. Histologically, it displays fibrous tissue with randomly oriented woven bone trabeculae. It is associated with a somatic mutation in the GNAS1 gene, encoding a stimulatory Gs-α protein. Liposclerosing myxofibrous tumor can greatly resemble FD, but Ragsdale²  asserts that the variety of tissue types is beyond the histologic spectrum of normal FD, although other authors⁵  disagree.

Intraosseous lipoma, a rare bone tumor, is discovered incidentally or after reports of pain.¹²  These lesions predominantly occur in the lower limb, usually within the calcaneus, but can also occur in the femur. Radiographically, these tumors present as a cystic lesion with increased radiolucency and a sclerotic rim. As the name would suggest, these lesions have a predominantly lipomatous component. Stage I lesions contain mostly mature lipocytes; however, higher-stage lesions can display foci of fat necrosis, cystic spaces, and dystrophic calcification.¹²  It is in this higher stage that intraosseous lipoma and LSMFT can become confused; LSMFT can also show lipomatous areas with involutional and ischemic change. However, intraosseous lipoma remains a very rare lesion, and the predominance of fat, with the absence of other histologic components such as fibromyxoid tissue, can distinguish it from LSMFT, although the age spectrum is similar to that of LSMFT.

Nonossifying fibroma is one of the most common benign fibrous lesions. Occurring mainly in adolescents and young adults, this tumor presents with symptoms of pain or can be asymptomatic.¹³  This tumor mainly occurs in the distal femur/proximal tibia, appearing as a radiolucent area with a sclerotic rim on radiographic imaging. Histologically, it is composed of fibrous tissue in a whorled or storiform pattern, often with admixed giant cells. This lesion can present with a similar radiographic image to that of LSMFT, but its occurrence in the lower femur and in young adults distinguishes it from LSMFT. The benign fibrous tissue pattern also does not include the variety of components seen with LSMFT.

In summary, the histologic features of LSMFT may suggest a variety of entities for consideration in the differential diagnosis. A comparison of the important features of these entities is featured in the Table.

Currently, the treatment of choice for this tumor depends on the clinical scenario. Asymptomatic lesions may not require specific therapy.¹⁴  Symptomatic lesions are usually treated with complete curettage and bone graft.^4,14  Patients presenting with pathologic fracture usually require joint arthroplasty. Typically, patients have a favorable prognosis after complete resection of tumor. However, several reports in the literature suggest a risk of malignant transformation.^2,3  Some authors report a risk as high as 10%, with malignancies such as osteosarcoma and malignant fibrous histiocytoma identified in a background of LSMFT.^3,7  This may be an overestimation, as some of these malignancies could have arisen de novo or in the background of other benign tumors.^4,15 

In conclusion, LSMFT is a rare, not yet universally recognized (ie, controversial) pathologic entity. Although it exhibits characteristic radiographic features, the variety of histologic patterns can create a unique diagnostic challenge for the pathologist. Because of its heterogeneous nature and potential histologic overlap with a variety of lesions, such as FD and intraosseous lipoma, and because its long-term behavior appears to differ somewhat from that of these other entities, LSMFT should remain an entity on the differential diagnosis for benign fibro-osseous lesions, especially when seen in the proximal femur. Further prospective study of this area, with more follow-up of well-classified lesions, also appears warranted.