ROM-based patient-specific structural analysis of vertebrae affected by metastasis

Prostate and breast cancer, the most prevalent kinds of cancer in Spain, present a high risk of developing spinal metastases, with a prevalence of 39%. The structural feasibility of the vertebrae is undermined by the tumors and there is a surge in the risk of collapse, leading to potential injury to the spinal cord. Hence, the study of the structural behavior of vertebrae with tumors is a research field with great potential to improve patients' overall living conditions. In this work, we propose to run personalized structural analyses of vertebrae with tumors using the Cartesian grid FEM (cgFEM), a non-geometry-conforming FE methodology capable of automatically creating a patient-specific model considering all the voxel-wise information of a CT scan. The use of these techniques in the highly demanding clinical environment requires obtaining the results of the structural analyses in real-time. For this, we propose a methodology to create a patient-specific reduced order model capable of evaluating the results of the structural analysis of a vertebra with a spherical tumor of any size, position and density in real-time, offering results very close to those that would be obtained by cgFEM. This would make possible analyzing a large set of tumor growth scenarios during the clinical process, providing valuable information to assess the risk of collapse and to define the most appropriate treatment for the patient.