Radiation induced sarcoma of bone is a rare but challenging disease process associated with a poor prognosis. To date, series are limited by small patient numbers; data to inform prognosis and the optimal management for these patients is needed. We hypothesized that patients with radiation-induced pelvic bone sarcomas would have worse surgical, oncologic, and functional outcomes than patients diagnosed with primary pelvic bone sarcomas

This was a multi-institution, comparative cohort analysis. A retrospective chart review was performed of all patients diagnosed with a radiation-induced pelvic and sacral bone sarcoma between January 1st, 1985 and January 1st, 2020 (defined as a histologically confirmed bone sarcoma of the pelvis in a previously irradiated field with a minimum 3-year interval between radiation and sarcoma diagnosis). We also identified a comparison group including all patients diagnosed with a primary pelvic osteosarcoma/spindle cell sarcoma of bone (i.e. eligible for osteosarcoma-type chemotherapy) during the same time interval. The primary outcome measure was disease-free and overall survival.

We identified 85 patients with primary osteosarcoma of the pelvis (POP) and 39 patients with confirmed radiation induced sarcoma of the bony pelvis (RISB) undergoing surgical resection. Patients with RISB were older than patients with POP (50.5 years vs. 36.5 years, p67.7% of patients with POP underwent limb salvage as compared to 77% of patients with RISB; the type of surgery was not different between groups (p=.0.24). There was no difference in the rate of margin positive surgery for RISB vs. POP (21.1% vs. 14.1%, p=0.16). For patients undergoing surgical resection, the rate of surgical complications was high, with more RISB patients experiencing complications (79.5%) than POP patients (64.7%); this approached statistical significance (p=0.09).

15.4% of patients with RISB died perioperative period (within 90 days of surgery) as compared to 3.5% of patients with POP (p= 0.02). For patients undergoing surgical resection, 5-year OS was significantly worse for patients with RISB vs. POP (27.3% vs. 47.7%, p=0.02). When considering only patients without metastatic disease at presentation, a significant difference in 5-year survival remains for patients with RISB vs. POP (28.6% vs. 50%, p=0.03) was a trend towards poorer 5-year DFS for patients with RISB vs. POP (30% vs. 47.5%), though this did not achieve statistical significance (p=0.09).

POP and RISB represent challenging disease processes and the oncologic outcomes are similarly poor between the two; however, the disease course for patients with RISB appears to be worse overall. While surgery can result in a favorable outcome for a small subset of patients, surgical treatment is fraught with complications.

The rate of fracture and subsequent nonunion after radiation therapy for soft-tissue sarcomas and bone tumors has been demonstrated to quite high. There is a paucity of data describing the optimal treatment for these nonunions. Free vascularized fibular grafts (FVFG) have been used successfully in the treatment of large segmental bone defects in the axial and appendicular skeleton, however, their efficacy with respect to treatment of radiated nonunions remains unclear. The purpose of the study was to assess the 1) union rate, 2) clinical outcomes, and 3) complications following FVFG for radiation-induced femoral fracture nonunions.

We identified 24 patients who underwent FVFG for the treatment of radiation-induced femoral fracture nonunion between 1991 and 2015. Medical records were reviewed in order to determine oncologic diagnosis, total preoperative radiation dose, type of surgical treatment for the nonunion, clinical outcomes, and postoperative complications. There were 11 males and 13 females, with a mean age of 59 years (range, 29 – 78) and a mean follow-up duration of 61 months (range, 10 – 183 months). Three patients had a history of diabetes mellitus and three were current tobacco users at the time of FVFG. No patient was receiving chemotherapy during recovery from FVFG. Oncologic diagnoses included unspecified soft tissue sarcomas (n = 5), undifferentiated pleomorphic sarcoma (UPS) (n = 3), myxofibrosarcoma (n = 3), liposarcoma (n = 2), Ewing's sarcoma (n = 2), lymphoma (n = 2), hemangiopericytoma, leiomyosarcoma, multiple myeloma, myxoid chondrosarcoma, myxoid liposarcoma, neurofibrosarcoma, and renal cell carcinoma.

Mean total radiation dose was 56.3 Gy (range, 39 – 72.5), given at a mean of 10.2 years prior to FVFG. The average FVFG length was 16.4 cm. In addition to FVFG, 13 patients underwent simultaneous autogenous iliac crest bone grafting, nine had other cancellous autografting, one received cancellous allograft, and three were treated with synthetic graft products. The FVFG was fixed as an onlay graft using lag screws in all cases, additional fixation was obtained with an intramedullary nail (n = 19), dynamic compression plate (n = 2), blade plate (n = 2), or lateral locking plate (n = 1).

Nineteen (79%) fractures went on to union at a mean of 13.1 months (range, 4.8 – 28.1 months). Musculoskeletal Tumor Society scores improved from eight preoperatively to 22 at latest follow-up (p < 0.0001). Among the five fractures that failed to unite, two were converted to proximal femoral replacements (PFR), two remained stable pseudarthroses, and one was converted to a total hip arthroplasty. A 6th case did unite initially, however, subsequent failure lead to PFR. Seven patients (29%) required a second operative grafting. There were five additional complications including three infections, one wound dehiscence, and one screw fracture. No patient required amputation.

Free vascularized fibular grafts are a reliable treatment option for radiation-induced pathologic femoral fracture nonunions, providing a union rate of 79%. Surgeons should remain cognizant, however, of the elevated rate of infectious complications and need for additional operative grafting procedures.

This text has been removed at the authors' request.