BACKGROUND: Denosumab is an inhibitor of monoclonal receptor activator of nuclear factor-ĸB ligand, approved to treat giant cell tumors of bone (GCTB). It is commonly used for unresectable tumors and for downstaging the tumor to perform less-morbid procedures. Although denosumab has been used extensively for GCTBs, there are no recommendations regarding the duration of therapy. The risk factors associated with local recurrence (LR) in patients receiving preoperative denosumab for GCTB also are unknown.
QUESTIONS/PURPOSES: (1) Is short-course (three doses or fewer) preoperative denosumab treatment as effective as longer course (more than three doses) of treatment in terms of achieving a clinical, radiologic, and histologic response in patients with GCTB? (2) Is there an increased risk of LR after short-course denosumab therapy compared with long-course denosumab therapy; and after controlling for confounding variables, what factors were associated with LR after surgery for GCTB in patients receiving preoperative denosumab?
METHODS: A retrospective study was performed using an institutional database of 161 skeletally mature patients with a histologic diagnosis of GCTB who received denosumab between November 2010 and July 2019 to downstage the tumor before surgery. In general, we used denosumab when we thought it would facilitate either resection or curettage (by formation of a sclerotic rim around the osteolytic lesion), when a less-morbid procedure than initially planned might be performed, and in patients with complex presentations like cortical breech and soft tissue extension, pathological fracture, thinning of more than three cortices of the extremity. From 2010 to late 2015, denosumab was administered for approximately 4 to 6 months; starting in late 2015 through 2020, the number of denosumab doses has been reduced. We divided patients into two groups: Those who received three or fewer doses of denosumab (short-course, n = 98) and those who received more than three doses of denosumab (long-course, n = 63). Comparing those in the long-course group with those in the short-course group whose procedures were performed at least 2 years ago, there were no differences in loss to follow-up before 2 years (3% [3 of 98] versus. 3% [2 of 63]). The mean patient age was 30 years (± 6.1) and the mean number of denosumab doses was 4.4 (range 1 to 14). Overall, 77% (37 of 48) of patients taking short-course denosumab and 75% (27 of 36) of patients on long-course denosumab underwent curettage, and the remaining patients with an inadequate bony shell around the tumor or destruction of articular cartilage in both groups underwent tumor resection. With the numbers available, the patients with short- and long-course denosumab were not different in terms of age, sex, MSTS score on presentation, lesion size, lesion location, Campanacci grade, presence of pathological fracture and pulmonary metastasis on presentation, and the type of surgery performed (curettage versus resection). We analyzed the change in the Musculoskeletal Tumor Society score, change in Campanacci grade, radiologic objective tumor response (defined as a partial or complete response, per the modified inverse Choi criteria), and histologic response (defined as reduction of more than 90% of osteoclast-like giant cells or a reduction of more than 50% of mesenchymal spindle-like stromal cells, along with evidence of lamellar or woven bone formation, when compared with the biopsy sample) between the two groups (short- and long-course denosumab). LR rates were compared between the two groups, and after controlling for confounding variables, factors associated with LR in all operated patients were analyzed with a Cox proportional hazards regression analysis.
RESULTS: With the numbers available, there was no difference between the short- and long-course denosumab groups in terms of mean percentage improvement in MSTS score (20 [± 18.5] versus 24 [± 12.6]; p = 0.37), radiologic objective tumor response (90% [43 of 48] versus 81% [29 of 36]; p = 0.24) and histologic response (79% [38 of 48] versus 83% [30 of 36]; p = 0.81). With the numbers available, there was no difference between the short- and long-course denosumab groups in terms of Kaplan-Meier survivorship free from LR at 5 years after surgery (73% [95% confidence interval, 68 to 76] versus 64% [95% CI 59 to 68]; log-rank p = 0.50). After controlling for potential confounding variables like age, sex, Campanacci grade and MSTS score on presentation, number of denosumab doses administered before surgery, clinical, radiologic and histologic response to denosumab, and time duration between denosumab therapy and surgery, we found that tumors involving the bones of the hand and the foot (hazard ratio 7.4 [95% CI 2.0 to 27.3]; p = 0.009) and curettage (HR 6.4 [95% CI 2.8 to 23.0]; p = 0.037) were independently associated with a higher risk of LR.
CONCLUSIONS: In this preliminary, single-center study, we found that a short-course of preoperative denosumab (three or fewer doses) was associated with no differences in clinical scores, histological and radiological response, or LR-free survivorship, compared with longer-course of denosumab (more than three doses). Fewer preoperative doses can reduce the complications and costs associated with more-prolonged therapy. Denosumab must be used cautiously before curettage for GCTB, and only if the benefit of joint salvage outweighs the possibility of LR. However, given the small number of patients, potentially clinically important differences might have been missed, and so our findings need to be confirmed by larger, multicenter, prospective trials.
LEVEL OF EVIDENCE: Level III, therapeutic study.