Multivariable statistical classification of 3D bone microarchitecture using morphological and mechanical features

Evaluating trabecular bone microarchitecture is not possible in the clinical routine nowadays. Based on simplified skeleton and Finite Element models, we have developed different image processing and simulation techniques to investigate bone microarchitecture and its mechanical stiffness. This work defines improved models for 3D bone architecture that generates multiple features. These parameters can be combined in a statistical discriminant analysis in order to study bone diseases such as osteoporosis. A clinical study is led on 2 populations of arthritic and osteoporotic bone samples. The results show the ability of our improved techniques to discriminate the 2 populations.