Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Many bone tumors contain large numbers of osteoclast-like giant cells. However, two lesions have many overlapping histologic findings and may pose a particularly difficult problem in differential diagnosis. These lesions are conventional giant cell tumor of bone and giant cell reparative granuloma. As in all bone pathology, the distinction cannot be made solely on histologic grounds. Careful considerations of conventional radiographic features as well as the clinical presentation are necessary for accurate interpretation.
Giant cell tumor of bone, sometimes referred to as conventional giant cell tumor, is a benign, locally aggressive neoplasm that usually affects young adults; about two thirds of patients are between the ages of 20 and 40 years. Giant cell tumors may occasionally occur in the elderly—a patient age 74 years has been documented. However, these neoplasms are extremely rare in growing children. Less than 2% of giant cell tumors affect patients younger than age 15 years.
Giant cell tumors occur most commonly in the distal femur, proximal tibia, and distal radius. These locations account for about 65% of cases. Other common locations include the pelvis, vertebral bodies, and proximal femur. Patients almost always present with pain, and a few present with a pathologic fracture. On rare occasions, giant cell tumor may be multicentric; as many as nine foci may occur either synchronously or metachronously.
Giant cell tumors, when they occur in the long bones, have a diagnostic radiographic pattern. A well-defined lytic lesion involves both the epiphysis and the metaphysis and almost always extends to the subchondral bone ( Fig. 11-1 ). The epiphyseal plate is almost always closed. Although well circumscribed, giant cell tumors usually lack a sclerotic rim. At least one cortex is thin and may also be expanded or destroyed ( Fig. 11-2 ). In the flat bones, giant cell tumors are also well-defined lytic lesions without a sclerotic rim.
Histologically, giant cell tumors consist of multinucleated giant cells admixed with mononuclear stromal cells ( Fig. 11-3 ). The stromal cells are polygonal or slightly elongated ( Fig. 11-4 ). Mitotic figures in the stromal cells are numerous. The multinucleated giant cells, often numerous, resemble osteoclasts. Formerly, particularly in the British literature, giant cell tumors were called osteoclastomas due to the abundance of these cells. Giant cell tumor, as well as chondroblastoma, harbor somatic driver mutations in the H3F3A and/or H3F3B genes. Mature cartilage or chondroid matrix is not present in giant cell tumors.
Immunohistochemical studies of giant cell tumors have clarified the relationship of the stromal cells to the giant cells. These studies suggest that there are two populations of stromal cells. One population, thought to be the neoplastic component, consists of spindle-shaped cells. The other population consist of polygonal cells that resemble macrophages. These two populations are immunohistochemically distinct; the polygonal cells stain for macrophage-associated antigens, particularly CD11a, CD18, and CD13, whereas the spindle cells do not. Many antigen profiles have supported this distinction. However, investigators have found this differential staining to be most striking with the CD68 stain. With this stain, the giant cells stain identically to the macrophage-like cells, which strongly suggests that the giant cells originate from fusion of the macrophage-like cells and not the neoplastic spindle cells. Giant cells also stain for acid phosphatase, a reaction also exhibited by bone-resorbing osteoclasts. These histochemical observations suggest that giant cells are of macrophage origin and are very similar to true osteoclasts. However, they do not result from fusion of the neoplastic cells. The stromal cells of giant cell tumor also have been shown to express p63.
In addition to the typical histologic features of giant cells and stromal cells, giant cell tumors frequently undergo secondary histologic changes, and these changes often lead to diagnostic confusion ( Fig. 11-5 ). First, focal necrosis is common; only the ghosts of the stromal cells and giant cells may be seen. On rare occasions, an entire neoplasm may be necrotic. Second, giant cell tumors may have areas of fibrohistiocytic reparative tissue ( Fig. 11-6 ). Spindle cells in a storiform pattern may be mixed with foam cells. In these areas, reactive bone formation is common. Senescent giant cell tumors may be composed entirely of this reparative tissue, and diagnostic giant cells and stromal cells may be minimal or absent. Probably, many examples of the entity “benign fibrous histiocytoma of bone” are ancient giant cell tumors that have been effaced by this fibrohistiocytic reaction. Finally, giant cell tumors may have focal areas containing an aneurysmal bone cyst, which may obscure the giant cells and stroma. If these secondary changes—necrosis, fibrosis with reactive bone, or aneurysmal bone cyst—predominate, the pathologist must rely on characteristic radiographic features to make the diagnosis of giant cell tumor.
Attempts to identify prognostic features, particularly the likelihood of recurrence, have met with variable success. Formerly, cytologic grading of stromal cells was thought to predict behavior of giant cell tumors. However, this practice is subjective and does not correlate with recurrence. Furthermore, the proliferation index and vascular density of giant cell tumors does not correlate with prognosis. Recently, a histomorphometric study of giant cell tumors has identified some cellular features that correlate with aggressive behavior. However, this cumbersome tool is not clinically practical. Other investigators have studied telomerase activity, metalloproteinase expression, and transforming growth factor-β in giant cell tumors. These studies may contribute to understanding the behavior of this neoplasm.
Some benign giant cell tumors have a deceptively ominous appearance. Lesions may have a very cellular stroma, and the cells contain plump, slightly hyperchromatic nuclei. Mitotic figures may be numerous. In addition, many giant cell tumors, 40% in some series, show evidence of vascular invasion. Plugs of both stromal cells and giant cells are present in vascular lumens adjacent to the neoplasm ( Fig. 11-7 ). Vascular invasion is not associated with a worse prognosis, although it may be the mechanism of “benign metastasis.”
A rare behavior of conventional giant cell tumor, present in 4% of cases, is pulmonary metastasis, the so-called “benign metastasizing giant cell tumor.” The pulmonary metastases are usually detected a few years after surgery and probably result from tumor embolization. In these unusual cases, the primary bone lesion is identical in all clinical, radiologic, and histologic respects to other conventional giant cell tumors. Also, the pulmonary metastases are histologically benign. This complication does not indicate an unfavorable prognosis. Pulmonary lesions grow very slowly and may be surgically resected. In fact, some pulmonary lesions remain stationary or even resolve. However, approximately 10% of patients with these “benign” lung metastases die of their disease.
Malignant giant cell tumor is a high-grade sarcoma developing in association with conventional giant cell tumor. This combination, when identifiable at the time of the initial presentation, is known as primary malignant giant cell tumor. Alternatively, secondary malignant giant cell tumor is the late development of a sarcoma after radiotherapy or a curettage of a giant cell tumor. Both manifestations of malignant giant cell tumor are extremely rare and account for about 7% of all giant cell tumors. Primary malignant giant cell tumors probably are manifestations of dedifferentiation, similar to that seen in parosteal osteosarcoma or chordomas. By contrast, secondary sarcomatous development in an irradiated giant cell tumor may be induced by radiation. The risk of malignant transformation in an irradiated giant cell tumor averages 20%. Reported time intervals until the development of sarcoma have been 4 to 39 years. As expected, patients with secondary malignant giant cell tumors are older than those with conventional giant cell tumors, the peak incidence being in the fifth decade of life. Both primary and secondary malignant giant cell tumors develop in locations typical of conventional giant cell tumor.
Malignant giant cell tumors, both primary and secondary, occur in the same radiographic setting as the conventional variant—the epiphyseal end of a long bone ( Fig. 11-8 ). The radiographic pattern of a primary malignant giant cell tumor may be identical to a conventional lesion. Most, however, are poorly circumscribed. A large volume of sarcomatous tissue may cause significant cortical destruction and soft tissue invasion. By contrast, secondary malignant giant cell tumors, occurring years after the original lesion, have a more destructive pattern. The location at the end of a long bone may be the only clue to its relationship to the original conventional giant cell tumor.
Histologically, primary malignant giant cell tumors are biphasic neoplasms; islands of conventional giant cell tumor are juxtaposed to a high-grade sarcoma. The sarcomatous component is usually a malignant fibrous histiocytoma, although occasionally an osteosarcoma is present. The sarcomatous component may comprise less than 20% of the volume of the neoplasm, emphasizing that conventional giant cell tumors must be generously sampled for histologic study.
Although 95% of conventional giant cell tumors never metastasize, they are very difficult to treat because of their local invasiveness. In the past, treatment with curettage and bone grafting led to a local recurrence rate of 30% to 40%. In the past 10 years, the treatment strategy has evolved to include wide exteriorization of the cortex, curettage with hand and power instruments, chemical cauterization with phenol, and reconstruction with subchondral cancellous bone grafts and methyl methacrylate. Internal fixation is often necessary. Wide exteriorization involves removing all the cortical bone over the lesion so that the surgeon does not have to work around any corners during curettage. The gross lesional tissue is removed with hand curettes, and then the remaining bone is burred down with power instruments to extend the curettage. This procedure removes the small (1- to 2-mm) pockets of the neoplasm. If the lesion has extended to within 1 cm of the articular cartilage, this area is reconstructed with use of a cancellous graft. The large remaining defect is filled with cement to prevent fracture and allow early motion. With this technique, which allows for very extensive curettage, the local failure rate is about 25%. The risk of arthritis is small if the cement is placed greater than 1 cm from the articular surfaces. If a giant cell tumor occurs in a location where wide excision is possible, such as the distal ulna, the recurrence rate is less than 5%. For recurrent lesions, recurettage is the treatment of choice.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here