Assessing activity-specific fracture risk in femora with metastatic lesions

Background: Metastatic bone disease involving the femur can lead to spontaneous fractures during daily activities. Activity-specific fracture risk prediction, for example during gait, could help to preserve patients' quality of life. While Finite Element (FE) models have been used under simplified conditions to assess fracture risk in intact femora, their accuracy for femora with metastatic lesions remains uncertain. Therefore, this study compares the fracture risk predictions of a simplified FE model, only including the peak hip joint load, with that of a physiological FE model, considering muscle- and hip joint forces of the whole gait cycle. Methodology: The simulations were conducted based on an intact femur Computed Tomography (CT) scan of a female (58 years). 93 variations of this CT scan were created by randomly inserting lytic lesions. For the physiological FE model, muscle- and hip joint forces were computed for a full gait cycle using a musculoskeletal model (OpenSim, Gait 2392). All 93 CT scans were converted into voxel-based nonlinear FE models with an elasto-plastic material, applying boundary conditions following literature. For the simplified FE models, the femora were shortened to approximately 30 % of their original length and only the peak hip joint load of the musculoskeletal simulation was applied. In both FE model approaches, bones were classified as “fractured” in case of a plastic displacement of the hip joint centre of at least 0.1 mm. Results: The fracture risk predictions of the two model types agreed in 90.32 % of cases. Prediction disagreement was predominantly associated with lesions located in the proximal femoral shaft or in the intertrochanteric region. Conclusion: Given the high level of agreement in fracture risk prediction between the simplified and physiological FE model during gait, it can be concluded that the simplified model offers a promising prognostic tool that combines predictive accuracy with clinical applicability. However, the study needs to be extended to verify these results for multiple subjects and different activities of daily living.