Bone strength prediction following tumor surgery using voxel-based finite element method validated by in-vitro mechanical tests

BackhroundsGiant cell tumor (GCT) mostely affects distal femur. After removing the tumor, the defectis reconstructed by cement infilling. Fracture is a frequent post-operative complicationthat limits the patients’ daily activities, and so a second operation might be likely needed.So, non-invasive methods for predicting the fracture risk is of great importance. To date,there is no firm biomechanical data to identify patients, affected by GCT and at high riskof fracture, for whom prophylactic actions should be made [1]. It is known that thereis a correlation between the risk of bone fracture and its strength [2], thus an accurateprediction of bone strength is needed, if one intends to estimate its fracture risk. In thisstudy, an experimentally validated finite element (FE) method, based on quantitative CT(QCT) images for predicting bone strength following tumor surgery, was introducedMethodsNine cadaveric distal femora were evaluated. GCT surgery was simulated on the femora byan orthopedic surgeon. Each femur was QCT scanned and then tested under destructivecompression load applied by a -25mm in diameter actuator to simulate stance-phaseloading (Fig. -1A). The maximum force achieved was considered as the bone strength(FTest). Three dimensional voxel-based FE models were created directly from convertingeach voxel of the QCT images into a cubic element (Fig -1B). Non-linear elastic-plastic(Fig. -1C) material properties were assigned to each element according to experimentallyderived relations based on its density value [2]. Cement was considered as a homogenousmaterial. Boundary conditions for FE models mimicked those applied in the mechanical tests. The proximal end of the model was fixed and displacement was applied to nodes in a-25mm in diameter circle region located on the medial condyle (Fig. -1B).Sum of the reaction forces of these nodes, versus their displacement along the loaddirection was drawn, and the maximum force of the graph was considered as bone strength(FFE). In order to evaluate, and validate the FE models, they were divided into two groups:5 femora in the development group (DG), and 4 femora in the evaluation group (EG). InDG, the material properties of the elements were modified until the difference betweenFFE and FTest approaches zero, employing a paired t test. In order to evaluate the accuracyof the predicted strength, the material properties employed in DG were assigned to EG’smodels, then results of FE analysis for EG were compared with experiments. Materialproperties modifications continued until the results of two groups, i.e DG and EG wereconsistent with the experimental data.ResultsVoxel-based FE models had 156 to 192 different material properties for bone (Fig.-1B), withmodulus of elasticity (E) and strength (S) of nearly zero to 26-21 GPa, and 175-140 MPa,respectively, depending on the specimen. The relationship between the bone strengthobtained by FE analysis (FFE) versus the strength found from in-vitro tests on cadavers(FTest) can be seen in Fig. 2, which shows a good agreement between FEM data andexperimentally measured of bone strength (R0.917= 2; p < 0.001(ConclusionsIdentifying patients at high risk of fractures prior to the surgery is very important. Currentclinical tools for identifying patients prone to fractures are limited to defect size andpatients’ pain, but they ignore many structural properties which are crucial in determiningbone strength. The QCT-based FE method presented here was capable of providing agood estimation of bone strength by considering material heterogeneity and post-failureproperties of bone as well as tumor geometry.References[1] A. Ghouchani, and G. Rouhi (2017), J. Med. Biol. Eng., 2[ ,67-454:)4(37] J. H. Keyak et al.(2005). Clin. Orthop. Relat. Res. 228-219 ;437