Patient-specific finite element modeling for fracture risk prediction in a canine model of osteosarcoma.

Abstract

In cancer patients with bone tumors, pathological fractures are a major concern. Making treatment decision for these patients requires an evaluation of fracture risk, which is currently based on semi-qualitative criteria that lack patient-specificity. Because of this, there exists a need for quantitative fracture risk prediction tailored to the patient's individual bone geometry. To address this need, this study aims to develop and validate a finite element (FE) technique that can be used to create patient-specific models and more accurately identify fracture risk. Model validation was performed using canine radii. Radii were harvested from eight canines euthanized for reasons unrelated to the study. A semicircular osteotomy was made in the distal portion of each bone to simulate tumor lysis. Samples underwent computed tomography (CT) scanning and were randomly assigned to loading groups for destructive mechanical testing. Three samples were tested in torsion, three in cantilever bending, and two in compression. FE models were created for each bone from the corresponding CT scan to replicate patient-specific geometry. Material properties were based on equations relating scan properties to elastic modulus. Boundary conditions and loads were added to the models based on the sample's treatment group. Stiffness and strain data were collected from both the mechanical testing and FE simulation, and yield load predictions were made based on maximum principal strain. Experimental and computational results were compared using a linear regression. The FE models were most accurate in predicting stiffness, followed by strain, with yield load having the lowest accuracy. Linear regressions resulted in R2 values of 0.9335 for bending and compression and 0.8798 for torsion. The proposed FE technique is a valid method for predicting fracture in a canine model of osteosarcoma. This method could provide patient-specific, quantitative data to aid clinicians in decisions regarding surgical intervention for patients with bone tumors.