Lytic spinal metastases are a big burden for cancer patients. These lesions are described as focal regions of very low bone mineral density (BMD), which cause a decrease in bone strength and an increase in the risk of fracture. The assessment of vertebral fracture risk in patients with spinal metastases is based on the Spinal Instability Neoplastic Score (SINS), which in many cases, is not able to provide a clear guidance. This problem is mainly due to the qualitative nature of the SINS, which therefore leads to a lack of objectivity in the assessment of patients with spinal metastases.

Finite element (FE) models have been extensively used to study the mechanical properties of healthy human vertebrae at different dimensional scales. FE models based on subjectspecific micro Computed Tomography (microCT) images have been validated and used to estimate how the local properties of bone tissues affect bone structure. Such models can be applied to better understand the effect of lytic lesions on the local and structural properties of human vertebrae. This was the aim of the first two studies presented in this thesis. In the first study, microFE models predictions of local and structural properties of vertebral bodies were validated. The validated microFE modelling method was then applied to study the effect of lytic lesions with different properties (size and location) on the local and structural properties of human vertebrae, from a feasibility study performed only for a small parametric sample. On the other hand, subject-specific Quantitative Computed Tomography (QCT) based FE models have been validated and used to predict the fracture risk of osteoporotic human vertebrae. Moreover, it has been shown that lytic lesions can be approximated to low BMD bone tissues. Therefore, these models can also be used to estimate the strength of vertebrae with lytic lesions. Thus, a third study included the development of a methodology to generate subject-specific QCT-based FE models of vertebrae with lytic lesions, and to assess their stability based on the physiological loads estimated from a spinal model. Such methodology was then applied to a cohort of 8 patients with lytic spinal metastases to provide a biomechanical analysis of vertebrae with lytic lesions to help in the assessment of the fracture risk.

To conclude, in this thesis two approaches were developed using subject-specific FE models of different dimensional scales, to provide biomechanical analyses of the effect of lytic lesions on human vertebrae. Both approaches can be used with the SINS to provide a more objective assessment of the risk of fracture of patients with lytic spinal metastases. Future work on the improvement of these approaches is important to make them more robust and helpful in clinics.