Spinal bone metastases represent a serious complication that may lead to vertebral fractures and other skeletal-related events (SREs), with a substantial negative impact on patient’s quality of life. Appropriate treatment selection is therefore essential to optimize outcomes and preserve quality of life, underscoring the need for predictive tools to anticipate metastatic progression and its structural consequences. In this study, we present a fully automated, patient-specific pipeline for vertebral structural analysis based on computed tomography (CT) imaging. First, machine learning (ML) methods are used to perform semantic segmentation of CT scans, eliminating the need for manual, patient-by-patient segmentation. Second, boundary conditions are defined to account for inter-patient physiological variability. A ML algorithm trained on musculoskeletal models is employed to predict patient-specific vertebral loads from relevant parameters. These predicted loads are then mapped onto the target vertebrae using the Coherent Point Drift (CPD) algorithm for point-set registration, enabling automated localization of boundary conditions. Finally, structural simulations are conducted using the Cartesian Grid Finite Element Method (cgFEM) to evaluate fracture risk under multiple tumour scenarios, including variations in tumour location, size, and density. Overall, the proposed methodology integrates deep learning–based segmentation, musculoskeletal load prediction, CPD registration, and cgFEM simulation to deliver an automated and personalized assessment of vertebral structural behaviour in the presence of metastases. The results indicate that this framework has the potential to support improved evaluation and treatment planning for patients with spinal bone metastases.