Denosumab de-escalation and clinical course
The median overall duration of denosumab treatment was 58 months (range, 33–118) (Figure 1). The interval of denosumab treatment was extended after a median of 12 months (range, 8–36), and the median treatment time for de-escalation was 46 (range, 9–75) months (Table 4). The median number of denosumab injections was 33 (range, 20–103), with median numbers of 17 (range, 8-37) injections for standard treatment and 12 (range, 4–33) injections for de-escalated treatment. Eight of the 9 patients had received de-escalated treatment at the latest follow-up.
All seven patients with spinal GCTB had stable disease following 8-weekly treatment for a median of 9 months (range, 6–43). Five patients proceeded to 12-weekly treatment, while 4 patients showed stable disease following 12-weekly treatment for a median of 28 months (range, 16-36). One patient (Case 4) proceeded to 24-weekly treatment and showed stable disease for an additional 8 months. One patient (case 5) developed malignant transformation 2.3 years after the first denosumab treatment and three months after the first 12-weekly treatment. One patient (case 7) treated with de-escalated therapy from 8-weekly to 24-weekly treatment experienced tumor regrowth 6 months after de-escalation to 24-weekly treatment. The patient was re-treated with standard denosumab therapy. Imaging showed tumor reduction with sclerotic changes; hence, the patient received 8-weekly treatment of denosumab and demonstrated stable disease. One (case 8) of the two patients with femoral GCTB proceeded to 12-weekly treatment of denosumab, while the other patient (case 9) proceeded to 24-weekly treatment of denosumab, both of them had stable disease at the latest follow-up. All patients were alive at the latest follow-up visit. None of the patients, except for one with malignant transformation, developed pulmonary metastasis.
Plain radiography and Computed tomography (CT) revealed sclerotic changes inside and around the tumor in all 8 patients after receiving standard denosumab treatment. The median time to response (TTR) was 3.5 months (range, 1.1–5.3 months) (Table 5), and the median time to identify the maximum sclerotic change was 11 months (range, 6–17 months) after the first denosumab treatment. Sclerotic change was continuously identified on CT without progressive osteolytic change after an 8-weekly and a 12-weekly treatment in all nine patients. One patient achieved complete response (CR), while 8 achieved partial response (PR) following a standard denosumab treatment according to the MD Anderson (MDA) criteria, which remained completely unchanged during de-escalated treatment (Table 5). Two patients (cases 4 and 9) showed stable radiographic changes after 24-weekly treatment of denosumab, while one patient (case 7) developed recurrent tumor with progressive osteolytic change after 24-weekly treatment of denosumab, which demonstrated sclerotic change again after resuming standard treatment.
An extraskeletal mass was identified in five patients, with a median size of 8.3 cm (range, 2.6–15 cm) before denosumab treatment (Table 5). The extraskeletal mass was reduced to a median size of 7.0 cm (range, 0.2–11.8 cm) after standard treatment in all five patients, which was significantly smaller than that before treatment (P = 0.043). The median reduction rate of the extraskeletal mass was 21% (range, 2%–92%). The median time to identify initial tumor reduction was 4.6 (range, 1.1–5.3) months after the first denosumab treatment (Figure 2), while the median time to identify maximum tumor reduction was 14 (range, 3–17) months. Four patients showed stable disease following de-escalated treatment, while one patient (case 7) experienced tumor regrowth.
Imaging of spinal GCTB prior to denosumab treatment demonstrated <50% collapse in 2 patients, no collapse with >50% body involvement in 3 patients, and no collapse with ≤50% body involvement in 2 patients according to spinal instability neoplastic score (SINS). At the latest follow-up, six patients did not experience further spinal collapse, while one patient with T10 disease (case 2) progressed from <50% collapse to >50% collapse. The SINS for spinal GCTB prior to denosumab treatment was a median of 8 (range, 5–10) points; two patients had stable disease, while five patients had potentially unstable disease based on their SINS. The SINS improved to a median of 4 (range, 1–10) points after standard treatment (P = 0.024); six patients had stable disease, while one patient had potentially unstable diseased based on their SINS. One patient demonstrated further improvement in SINS after de-escalated treatment. The SINS further improved to a median of 3 (range, 1–10) points at the latest follow-up (P = 0.026), which was not significantly different from that after the standard treatment (P = 0.317). The tumors in six patients were classified as stable, while that in one patient was classified as potentially unstable after receiving de-escalated treatment.
Eight of the nine patients experienced local pain prior to denosumab treatment, and the median numerical rating scale (NRS) score was 4 (range, 0–10) (Table 6). Seven patients became pain-free (NRS=0) after standard denosumab treatment, and the median NRS score was 0 (range, 0–2) (P = 0.001). Pain disappeared at a median of 97 (range, 7–181) days after the first denosumab treatment and after a median of five (range, 2–8) injections of standard treatment. One patient (case 1) experienced pain (NRS score =2) at the latest follow-up. Tumor recurrence (case 7) was identified after treatment de-escalation to 24-weekly treatment due to progressive severe pain (NRS=10), which subsequently disappeared (NRS=0) after resuming standard treatment.
All seven patients with spinal GCTB experienced numbness or hypoesthesia in the lower legs, while two patients had bladder and bowel dysfunction. Two patients were classified as having grade D, while five was classified as having grade E prior to denosumab treatment, based on the AISA impairment scale. Sensory disturbance improved in all seven patients after receiving standard treatment, although three had slight numbness during standard and de-escalated treatment. One of the two patients with bladder and bowel dysfunction prior to treatment completely recovered, while the other experienced persistent mild constitution symptoms after denosumab treatment. All seven patients were classified as having grade E based on the American Spinal Injury Association (ASIA) impairment scale after receiving de-escalated treatment.
Two patients experienced ONJ after the treatment was de-escalated to every 12 weeks. One patient (case 2) developed ONJ 3.3 years after the first denosumab treatment and 2.3 years after the initial de-escalation. The other patient (case 7) developed ONJ 7.8 years after the first denosumab treatment and 5.2 years after the initial de-escalation. Both patients were treated for ONJ without discontinuation of the denosumab treatment. Four patients had asymptomatic hypocalcemia (median: 8.6 mg/dl; range, 8.2–8.7 mg/dl) and did not require additional treatment. Hypocalcemia was identified transiently at a median of 1 week (range, 1-2 weeks) after the initial denosumab treatment and was not detected thereafter. One patient with sacral GCTB (case 5) developed malignant transformation after 12-weekly treatment of denosumab. The patient had undergone primary surgical resection and developed local recurrence 4 months postoperatively. Multiple courses of selective embolization were performed as treatment for recurrent tumor, which remained stable for over 6 years. Atypical femoral fractures did not occur during denosumab treatment.
A 25-year-old woman with sacral GCTB presented with severe pain (NRS 9) and numbness in the left buttocks and lower limbs. A huge mass in the sacrum with cortical destruction and extensive soft tissue involvement was observed on CT images (Figure 3A). Magnetic resonance imaging (MRI) revealed a large tumor in the sacrum with high intensity on T2-weighted images (Figure 3B). The patient underwent standard denosumab therapy. The pain decreased to NRS 2 after one week and disappeared after two weeks. Five months later, MRI revealed remarkable shrinkage of the tumor (Figure 3C). The patient showed stable disease over the next two years, with remarkable sclerosis of the lytic lesions (Figure 3D). After 8-weekly treatment of denosumab for 6 months, the patient achieved stable disease (Figure 3E). She received denosumab therapy every 12 weeks. At the last follow-up, the patient showed a stable disease (Figure 3F, G).
A 26-year-old woman had a GCTB of the femoral head. CT images (Figure 4A) showed a lytic lesion. She underwent curettage and bone grafting with internal fixation, which were performed twice. CT examination revealed a lytic lesion in the femoral neck (Figure 4B). Hence, denosumab was initiated. After 4 months, a change in sclerosis was observed (Figure 4C). After 10 months, the lytic lesion was completely filled with newly formed bone (Figure 4D). Denosumab was injected every 24 weeks. At the last follow-up, the patient showed stable disease (Figure 4E).