Page 126 - Reconstructive Techniques in Musculoskeletal Tumor Surgery

Chapter 6
of the allograft during the primary surgery. Fractures were treated conservatively or with internal (n = 9) or external (n = 1) xation; all healed uneventfully. Of the 20 patients with a host bone fracture, 17 had had a reconstruction of ≥8 cm and four had had plate xation. In univariable analysis, reconstruction length of ≥8 cm (OR = 5.5), nonunion (OR = 9.8), and the extent of cortical resection signi cantly in uenced the risk of host bone fracture (table 2). In multivariable analysis, nonunion and the extent of resection retained signi cance (table 3).
Nonunion occurred in eight patients (7%). Five (5%) underwent revision of the osteosynthesis, combined with allogeneic bone grafting (n = 2), allogeneic and autologous bone grafting (n = 2), or tibial autografting (n = 1). Nonunion resulted in graft failure in two of the patients, ve and 24 months after the index procedure. Of the eight patients with nonunion, two received chemotherapy (p = 0.20) and one had radiation therapy (p = 0.38). The nonunion risk was higher for reconstructions of ≥8 cm in length (OR = 5.9) but this was not a signi cant factor (table 2).
Deep infection developed after eight reconstructions (7%), ve in the tibia, two in the femur, and one in the radius. Three infections (3%) were eradicated with surgical debridement and antibiotics, and the other ve resulted in graft failure (5%): two within the rst postoperative month and one each after eight, 33, and 34 months. The mean duration of surgery for the patients with an infection was 3.9 hours (SD = 3.6 hours) compared with 2.9 hours (SD = 1.5 hours) for those without an infection (p = 0.10). Reconstructions of the tibial diaphysis (OR = 4.2) and those comprising >50% of the cortical circumference (OR = 9.8) were associated with a signi cantly higher risk of infection (table 2).
Oncological Outcome
The margins obtained during excision of the eleven benign lesions were adequate in seven, questionable in two (one patient had additional cryosurgery), and intralesional in two (one patient had cryosurgery and one had phenolization), but clear margins were not the aim in all patients.
Of the 97 patients with a malignant lesion, ten (10%) had questionable margins and ten (10%) had an intralesional resection (table 4). The rates of inadequate margins were comparable among the grades of malignancy (p = 0.36). All computer-navigated resections resulted in adequate osseous margins, but there was one contaminated soft-tissue margin. Residual or recurrent tumor was diagnosed in 15 (15%) of the 97 patients with a malignant tumor, after a median of 12 months (1 day to 13 years). Of the 61 patients with a low-grade malignant tumor,
124