Case 1. A twenty-nine-year-old man with a history of GCT (6.3 × 5.6 × 3.2 cm) that was resected from the proximal part of the right fibula two years prior returned with a lateral knee mass. Radiographs demonstrated an expansile, lytic mass (5.5 × 3.0 × 2.7 cm) in the previous tumor site. Positron emission tomography (PET)/computed tomography (CT) highlighted this lesion and additionally demonstrated a separate hypermetabolic lesion (3.6 × 2.5 × 2.3 cm) within the soft tissue of the ipsilateral distal part of the thigh (Fig. 1). Based on its location and distance from the previous surgical site, this lesion was considered consistent with a metastasis. Biopsy of both lesions showed a monotonous population of small, bland epithelioid and spindle cells with diffusely present giant cells, confirming the diagnosis of GCT (Fig. 2).
As a participant in a phase-II open-label study, the patient received loading doses of denosumab on days eight and fifteen, followed by monthly 120-mg doses. Repeat imaging three months later confirmed a noteworthy response in both the fibula and thigh masses, with decreased size evident by CT and decreased metabolic activity evident by PET. Because of discomfort, the thigh lesion was resected six months after the start of treatment. Follow-up radiographic evaluation of the bone lesion confirmed a stable mass with no concerning features of GCT.
Grossly, the soft-tissue resection specimen consisted of a hard, oval intramuscular mass. Histologically, the tumor was well circumscribed and contained a dense shell of compact mature bone surrounding a fibro-osseous lesion with trabeculae of variably mature osteoid (Figs. 3-A and 3-B). In most areas, the osteoid formed a complex web of intertwined trabeculae within scant fibrous tissue. Toward the center of the specimen, bland mononuclear spindle cells were identified between the trabeculae; very few osteoblast-like cells were seen rimming the osteoid. Multinucleated osteoclast-type giant cells were notably absent, as were the typical histiocyte-like stromal cells of GCT. Although soft-tissue recurrences of GCT may form immature bone at the periphery of the implant, the pattern of bone formation seen was different from the mature, reactive bone with osteoblastic rimming that is often found around such GCT lesions. Overall, the appearance was suggestive of intramembranous ossification, a reactive process, and/or an unusual form of osteoma. Immunohistochemistry was negative for CD68 and p63, which are markers typically present in GCT. Based on the histological and immunohistochemical findings, the GCT was considered to have undergone complete remission, and the diagnosis rendered was “residual osteocytic lesion.”
Case 2. A thirty-four-year-old man had an incidental finding of a lytic lesion of the proximal part of the tibia during the workup of a sports-related injury. A biopsy was performed, and a diagnosis of GCT was made. Following surgical excision, biannual radiographs were obtained. Eighteen months later, a suspicious lesion appeared at the prior tumor site. Magnetic resonance imaging (MRI) confirmed a complex mass (2.7 × 2.8 × 3.5 cm) arising within the proximal part of the tibia, extending laterally through the cortex into the soft tissue (Fig. 4). Biopsy confirmed recurrent GCT. At that time, the patient joined the denosumab trial. On follow-up MRI three months later, the mass was noted to have shrunk, and it was stable at six months. After the full course of treatment, the residual lesion was curetted, and internal fixation was performed.
Grossly, the specimen consisted of fragments of bone and soft tissue, with no solid mass identified. Histologically, there were areas of new bone formation along with sheets of stromal cells with a web of thin osteoid trabeculae and an absence of giant cells, similar to the patient in Case 1. Immunohistochemistry demonstrated no expression of CD68 and p63. This was deemed consistent with denosumab-treated GCT.
Denosumab, a monoclonal antibody targeting RANKL, represents a potential therapy in the treatment of certain instances of GCT8. Subcutaneous administration provides rapid and sustained suppression of bone turnover in patients with multiple myeloma and osteolytic bone disease, and in patients with breast and prostate cancer with bone metastases8. It has been postulated that inhibition of RANKL by denosumab in patients with GCT might inhibit bone destruction and eliminate giant cells. In this report, we describe two cases: (1) a denosumab-treated metastatic GCT, which resulted in a completely bland residual osteocytic lesion, and (2) a denosumab-treated recurrent GCT, which resulted in a bland residual osteocytic lesion with new bone formation.
During the phase-II study, these two patients received subcutaneous injections of denosumab (120 mg monthly) following loading doses on days eight and fifteen. After six months of therapy, both the soft-tissue metastasis of the GCT and the intraosseous recurrent lesion differentiated into a bland calcified fibro-osseous mass. The lesions exhibited no semblance to a traditional GCT; instead, they had a pattern reminiscent of intramembranous ossification, reaction to injury, and/or a variant of osteoid osteoma. The lesions exhibited partial zonation with a peripheral rim of bone, a region of relatively mature osteoid-forming trabeculae with occasional stromal fibroblasts and vessels, and a central portion with immature osteoid that formed a complex interconnecting web of thicker trabeculae and scantier stroma. The case of intraosseous recurrence also showed signs of new bone formation. There were neither giant cells nor typical histiocyte-like stromal cells of GCT identified. Moreover, the lesions exhibited complete loss of expression of markers traditionally found in the multinucleated and mononuclear cell fractions of GCT, namely CD68 and p639. In Case 1, the original fibular lesion was not rebiopsied after treatment; it remains unclear whether the bone was affected in a manner similar to that of the soft-tissue specimen or whether it simply became predominately invested within remodeled bone, much like in Case 2.
While these lesions simulated other bone-forming lesions, they differed architecturally. An osteoid osteoma has a peripheral rim of mature bone, but centrally contains a nidus of organized trabeculae rimmed with osteoblasts. In our cases, the rimming osteoblasts were frequently absent, and thick trabeculae of immature osteoid were located centrally. With mature bone, myositis ossificans shows a zonation pattern from outside, followed by organized and parallel seams of osseous trabeculae rimmed with osteoblasts as well as a central portion frequently containing a fasciitis-like proliferation of myofibroblasts. While the residual osteocytic lesion exhibited zonation, the osteoid trabeculae formed among the stromal cells were more disorganized and interconnected, and frequently lacked rimming osteoblasts; in Case 1, the central portion was occupied by a ramifying maze of immature osteoid. Because the morphology and immunophenotype of treated GCT differ from that of any heretofore recognized entity, we feel the neologism “residual osteocytic lesion” is warranted.
It is possible that, in a manner analogous to osteocyte formation in bone, denosumab induces differentiation of GCT via its interaction with RANKL. This theory is in keeping with a growing view that the neoplastic stromal cell fraction in GCT is related to osteoblasts10. This would explain why the first lesion was entirely osseous with variation in the maturity of the formed osteoid. Presumably, the periphery is more easily exposed to the inducing/differentiating agent while the center is less exposed. In some way, the antibody must be osteoclast-lytic, resulting in the complete absence of giant cells. Presently, it remains unclear whether the viable central spindle cells in the residual osteocytic lesion have the potential to recapitulate a GCT in the absence of ongoing therapy. It is also important to note that several adverse events have been reported (pain in an extremity and/or back pain and headache are the most common)8.
These two cases are an example of how targeted therapy may result in stable residual disease with a change in morphology. In gastrointestinal stromal tumors treated with imatinib, the outcome frequently is a lesion consisting of hyaline, myxoid, and/or cystic degeneration with necrosis, largely bereft of tumor cells; even rhabdomyomatous differentiation has been reported11,12. Likewise, histologic and immunophenotypic differences have been found following targeted therapy in lymphoma13. In the future, it is conceivable that other RANKL-rich giant cell lesions, such as giant-cell reparative granuloma, giant-cell osteosarcoma, and perhaps even aneurysmal bone cysts, may also benefit from this or related treatments14,15.
We report the novel histopathologic transformation of GCTs successfully treated with denosumab. Although our sample size was small, pathologists should be aware of this occurrence, and more investigation should be done on future cases to see if there are any additional changes. Targeted therapies may result in histologic and immunophenotypic differences in neoplasms following treatment; thus, recognition and description of such changes are essential to promote accurate communication among pathologists, clinicians, and surgeons. Pathologists should be aware of this phenomenon and add “residual osteocytic lesion” to the differential diagnosis of bone-forming tumors.