Strength Analysis Of Cancellous Bone Based On Topology And Morphology Of Porous Structure

Predicting the mechanical properties of cancellous bone is significance to the personalized design of bone scaffolds.The measurement of bone mineral density(BMD)is universally recognized as the standard for diagnosing osteoporosis currently,however,BMD only accounts for approximately 60%—70% of the variability in bone strength..Studies have shown that trabecular bone microstructure is also an important element affecting bone strength.In order to improve the prediction accuracy of cancellous bone mechanical properties,it is necessary to establish the relationship between cancellous bone porous structure parameters,bone density and bone mechanical properties.In this paper,the parameters of cancellous bone porous structure were studied based on micro-CT images,and a multi-parameter fitting predictive model of cancellous bone modulus was proposed,which provided a technological foundation for the personalized design of bone implants.The main research content of this thesis are as follows:1.The cancellous bone data set was obtained by three-dimensional reconstruction of CT images.Using imagej,mimics and fuzzy skeletonization algorithm,a series of cancellous bone structural parameters were obtained.2.The multiple linear regression model is applied to the prediction of bone trabecular modulus.By observing the collinearity diagnosis results and non-normalized coefficients of each parameter in multiple linear regression,four parameters for predicting the trabecular bone modulus are obtained: bone volume fraction,axial bone volume fraction,fractal dimension,surface area volume ratio.Multiple linear regression was carried out on these four parameters.Prediction formula of the the cancellous bone(size 3.9mm×3.9mm×3.9mm)based on multiple linear regression was established.The coefficient of determination of the prediction formula was 0.888.In addition,the prediction formula for the modulus of cancellous bone(size 7.8mm× 7.8mm × 7.8mm)at large size was calculated,and its coefficient of determination was 0.906.3.Neural network is applied to predict trabecular modulus.A trabecular modulus predictive model is established using BP neural network which coefficient of determination is0.8824.In addition,the K-means clustering algorithm is used to classify the trabecular bone data into three categories,and the BP neural network is used to predict the trabecular bone data of the three types of trabecular bone data respectively.The coefficient of determination of the new predictive model is 0.919.