Prediction of compressive strength of vertebral body with metastatic lesions based on quantitative computed tomography-based subject-specific finite element models

Pathologic vertebral fractures are a major complication in metastatic spine disease. However, current clinical scores, such as Spinal Instability Neoplastic Score (SINS), show limited predictive capability, particularly within the indeterminate range where most clinical uncertainty lies. This study aimed to develop and evaluate quantitative computed tomography (qCT)-based subject-specific finite element (SSFE) models to predict vertebral strength in presence of different metastatic lesion types. Twelve ex vivo human spine segments, each containing one metastatic (n=12) and one adjacent control vertebra (n=12), were scanned using qCT and calibrated using a calibration phantom. Homogenised nonlinear finite element models were developed with spatially heterogeneous, isotropic, density-dependent material properties and loaded under uniaxial compression corresponding to 1.9% apparent strain. Ultimate failure load, stiffness, and strain distributions were compared between metastatic and control vertebrae. Predicted failure load ranged from 0.2 kN to 6.2 kN (mean. {+/-} standard deviation: 1.8 {+/-} 1.6 kN metastatic; 1.7 {+/-} 1.5 kN control), with no statistically significant difference between groups (p > 0.05). Normalised failure load varied widely, reflecting lesion-specific mechanical heterogeneity. Lytic lesions generally weakened vertebrae, whereas mixed and blastic lesions occasionally enhanced strength, likely due to localised sclerosis or reactive bone formation. High compressive axial strains (greater than 0.019) were frequently concentrated near the endplates, particularly in lytic vertebrae. qCT-derived bone mineral density strongly correlated with failure load (R{superscript 2} = 0.74-0.77). These findings highlight the complexity of metastatic vertebral mechanics and demonstrate that qCT-based SSFE modelling provides a quantitative framework for assessing fracture risk, complementing conventional imaging-based tools.