Failure Prediction of Tumoral Bone with Osteolytic Lesion in Mice

Abstract

Pathological fractures due to bone metastases remain difficult to predict. The aim of this study is to assess whether a model taking into account tumor-specific geometry and mechanical properties improves assessment of bone failure, and to determine which criterion has to be taken into account to improve failure detection. To achieve this aim, an osteolytic mice model was considered. Tumoral cells were injected intra-tibially to induce a tumor in the bone. After six weeks, eight mice were sacrificed. Tomographic (μCT) images were obtained to build subject-specific finite element models. A compression test was performed on each tibia and used to assess the finite element models. Implementation of tumor geometry and mechanical properties did not provide better failure prediction in comparison to models based on μCT grey levels. The average difference with experiments reached respectively (23 ± 22% and 12 ± 7%). Considering a detection criterion based on the percentage difference between bone global ultimate load and bone local ultimate load (with a region of interest surrounding the tumor) allowed detection of all bones that experienced a partial failure. A next step will be to assess this failure criterion on human bones to help clinicians in decision-making.