Tumor-bearing frozen autografts have been widely used for reconstruction of bone defects caused by tumor resection. However, some patients undergo removal of the grafted bone due to surgical site infection, tumor recurrence, or fractures of the grafted bone. In this retrospective cohort study, predictive factors for graft survival were investigated in 123 patients who underwent reconstructions using a tumor-bearing frozen autograft after bone tumor resection of the extremities. To determine the independent predictors of graft survival, the association between various parameters and graft survival was investigated. The graft survival rates were 83.2% at 5 years and 70.2% at 10 years. Among the 123 frozen autografts, 25 (20.3%) were removed because of complications. In univariate analyses, male sex, BMI of ≥23.6, tibia, and chemotherapy were significantly associated with poor graft survival, whereas the pedicle/hemicortical freezing procedure was significantly associated with better graft survival. Multivariate analysis using the Cox proportional hazards regression model revealed that BMI of ≥23.6 (HR, 3.4; p = 0.005), tibia (HR, 2.3; p = 0.047), and freezing procedure (HR, 0.3; p = 0.016) were independently associated with graft survival. Based on the results, pedicle or hemicortical freezing techniques are recommended in cases where these techniques can be applied.
Keywords: bone tumor; frozen autograft; graft survival; predictor; reconstruction.