The patient presented at this institution in June 2009 with severe pain and disability of the left leg at the site of the previous excisions. Radiographs demonstrated a large lucency around the cement (Figure ​(Figure1).1). Concurrent staging revealed nodules on chest radiograph, and computerized tomography revealed two nodules in each lung, up to 2.8 cm in dimension (Figure ​(Figure2).2). In July 2009, the patient underwent bilateral thoracoscopy-assisted pulmonary resections. Frozen section revealed the pulmonary lesions to be histologically benign. As a result of the benign pulmonary histology, a repeat intralesional excision, cauterization with phenol and methacrylation with internal fixation of the proximal tibia was performed. Intraoperatively, the tumor was recurrent throughout (cephalad, caudad, medial and lateral to the cement) and bulging poster medially to the pes anserine tendons. Fibrous tissue was observed near the patellar ligament insertion. The post-operative course was unremarkable. The permanent paraffin sections confirmed the benign histology in the lung lesions; however, the tibial recurrence was frankly malignant. As a result, this patient received pre-operative chemotherapy given the poor prognosis [5]. Following the chemotherapy, he was taken to surgery, and an en-bloc resection of the upper tibia was performed. He was reconstructed with a modular oncology tibial and knee prosthesis. No residual tumor was seen in the resected specimen. Close follow-up, consisting of chest CT and tibial imaging every 3 months, is ongoing.

The gross specimen from the tibial excision consisted of fragments of pink-tan and red-brown tissue with pieces of cement and bone fragments. The metastatic tumor nodules resected from the lung (Figure ​(Figure4)4) grossly were well circumscribed mahogany nodules and histologically showed essentially identical morphology to the original tumor. Some areas of the recurrent tibial tumor also show classic histology of benign giant cell tumor of bone (Figure ​(Figure5A).5A). Immunostains distinguish between the monocyte lineage cells--both mononuclear cells and multinucleated giant cells, which are highlighted by the lineage specific markers CD4 and CD43 (Figure ​(Figure5B,5B, inset; the smaller, darker CD43+ cells are T cells) and the stromal spindle cell component which is negative. About 50% of the tibial tumor showed undifferentiated pleomorphic sarcoma having large spindled cells with vesicular nuclei, prominent nucleoli and scattered mitoses (up to more than 20 per 10 standard high power fields) and abundant apoptotic debris (Figure ​(Figure5C).5C). The benign component showed nuclear P63 staining in a subset of cells, as has been reported (Figure ​(Figure5D)5D) [6]. Also, Mib-1 staining is low in the benign component and significantly increased in the malignant component (Figures 5E, F).

In karyotypic analysis of giant cell tumors of bone, end-to-end fusions of various chromosomes, termed telomeric associations, are seen in most tumors in a subset of cells, and these have been localized by FISH studies to the CD68 negative spindled stromal component[13]. At lower frequency than telomeric associations, clonal chromosome gains, deletions and translocations can also be found in GCT of bone[14]. Translocations occur more frequently in tumors that have more telomeric associations, and they probably result from ensuing problems in separation of dicentric chromosomes during telophase[14]. However, cells with chromosomal abnormalities do not have enough of a growth advantage to dominate the tumor population, and analyses of many metaphases is required to recognize that there are clonal subpopulations in the tumor[14]. In Goronova's study[14] there was no association between karyotype and prognosis. Also, no dominant recurrent cytogenetic abnormality was seen that might give clues to pathogenesis. Thus, this peculiar low level chromosome instability reflects an underlying defect in chromosome maintenance, though probably not due to a generalized defect in telomerase activity[15]. The minor genomic instability seen in vitro is reflected by a low but definite probability of transformation clinically, as seen here. However, because of the current lack of a defined molecular or cytogenetic marker of GCT of bone, formal proof that the rare cases of malignant transformation are due to evolution of the spindle cells of GCT of bone is lacking. An alternative hypothesis would be that that GCT of bone predisposes to malignancy in an unrelated cell--akin to secondary malignancy in the setting of osteonecrosis or Paget's disease of bone. Notably, cases are considered secondary malignant giant cell tumors even if, on recurrence, only the malignant component is seen.

This case illustrates a sarcoma arising at the site of a locally recurrent GCT of bone, with concurrent non-sarcomatous metastases. Local recurrence of GCT of bone is common, occurring 10-20% of the time[16-18], and depends on the aggressiveness of the initial surgery[19]. Metastasis with "benign" morphology, most commonly to the lung, was reported in an average of 3% of cases (table ​(table1),1), and more than half occurred in patients who also had local recurrence.

Diagnosis of transformation relies on overt malignant cytological features, as necrosis and scattered mitoses can be seen in the usual "benign" giant cell tumor of bone. Immunohistochemical studies are not required for the diagnosis of transformation. Usually there is no evident lineage of maturation, but cases of osteosarcoma have been reported[9,20-24]. The incidence of malignancy in GCT varies considerably among studies, with an overall rate of approximately 5% considering the largest series together (table ​(table2).2). About half were associated with prior irradiation (table ​(table2),2), but transformation was found in the initial resection in about one third of cases (table ​(table2).2). Secondary transformation had a latency of up to 28 years[20] in a case that without radiotherapy. Given the wide range of reported incidence of malignancy in giant cell tumor of bone among studies, it is probable that this entity has been over diagnosed in the reported series. Of note, in the publications with the highest rates of malignancy, there was no pathological slide review. In Bertoni's study, pathological review was performed[20], and as a result, the initial diagnosis was revised to something other than malignant transformation of giant cell tumor in 12 of 26 cases initially considered for which slides were available. Anract's study, which did not include pathological review, includes 5 cases with "well differentiated fibrosarcoma"[5], which is much different from the more widely recognized high grade transformation. Likewise, in Domovitov's report, 57% of the cases had the curious diagnosis of "focally" malignant histology. Accordingly, the outcomes in their study were optimistic, with 80% of patients predicted to be recurrence free at 5 years based on their data, compared to the 10-50% overall survival estimated in most other series. Thus, excluding radiation-induced transformation cases, and considering only series with pathological review, the incidence of malignant transformation of giant cell tumor of bone is only about 1%, which is the rate cited in the WHO publication[25].