Objectives: To formulate a scoring system enabling decision making for prophylactic stabilization of the femur following surgical resection of a soft tissue sarcoma (STS) of the thigh.
Methods: A logistic regression model was developed using patient variables collected from a prospective database. The test group included 22 patients with radiation-related pathological femur fracture following surgery and radiation for a thigh STS. The control group of 79 patients had similar treatment but without a fracture. No patients received chemotherapy. Mean follow-up was 8.6 years. Variables examined were: Age (<
49, 50–70, >
70 years), gender, tumor size (0–7, 8–14, >
14 cm), radiation dose (low=5000 cGy, high>
6000 cGy), extent of periosteal stripping (<
10, 10–20, >
20 cm) and thigh compartment (posterior, adductor, anterior). A score was assigned to each variable category based on the coefficients obtained in the logistic regression model.
Results: Based on the regression model and an optimal cut-point, the ability to predict radiation associated fracture risk was 91% sensitive and 86% specific. The area under the Receiver Operating Characteristic (ROC) curve was 0.9, which supports this model as a very accurate predictor.
Conclusions: Radiation-related femur fractures following combined surgery and radiation treatment for STS are uncommon, but are difficult to manage and their non-union rate is extremely high. These results suggest that it is possible to predict radiation-associated pathological fracture risk with high sensitivity and specificity. This would allow identification of high risk patients and treatment with prophylactic IM nail stabilization. Presentation of this model as a clinical nomogram will facilitate its clinical use.
