Assessment of global morphological and topological changes in trabecular structure under the bone resorption process

Osteoporosis is a frequent skeletal disease characterised both by loss of bone mineral mass and deterioration of cancellous bone micro-architecture. It can be caused by mechanical disuse, estrogen deficiency or natural age-related resorption process. Numerical analysis of high-resolution images of the trabecular network is recognised as a powerful tool for assessment of structural characteristics. Using μCT images of 73 thoracic and 78 lumbar human vertebral specimens in vitro with isotropic resolution of 26μm we simulate bone atrophy as random resorption of bone surface voxels. Global morphological and topological characteristics provided by four Minkowski Functionals (MF) are calculated for two numerical resorption models with and without conservation of global topological connectivity of the trabecular network, which simulates different types of bone loss in osteoporosis, as it has been described in males and females. Diagnostic performance of morphological and topological characteristics as a function of relative bone loss is evaluated by a correlation analysis with respect to experimentally measured Maximum Compressive Strength (MCS). In both resorption models the second MF, which coincides with bone surface fraction BS/TV, demonstrates almost constant value of Pearson's correlation coefficient with respect to the relative bone loss ΔBV/TV. This morphological characteristic does not vary considerably under age-related random resorption and can be used for predicting bone strength in the elderly. The third and fourth MF demonstrate an increasing correlation coefficients with MCS after applying random bone surface thinning without preserving topological connectivity, what can be used for improvement of evaluation of the current state of the structure.