Finite element fracture load analysis and dark-field X-ray imaging of osteoporotic and healthy vertebrae in human lumbar spine specimens

Purpose: This study investigated the association of measurements from a clinical X-ray dark-field prototype system and CT-based finite element analysis (FEA) in lumbar spine specimens. Materials and Methods: In this prospective study, human cadaveric spine specimens (L2 to L4) were examined using a clinical prototype for dark-field radiography, yielding both attenuation and dark-field images. Specimens were scanned in vertical and horizontal positions. Volumetric bone mineral density (BMD) values were derived from quantitative CT measurements. Bone segmentation masks derived from CT-images were used for FEA-estimated fracture load (FL) calculations. FEA-estimated FL, dark-field, and attenuation signals were compared between osteoporotic/osteopenic (BMD < 120 mg/cm³) and non-osteoporotic/osteopenic specimens using the paired t-test and the Wilcoxon Mann–Whitney U test. Associations were tested using Spearman correlation. Results: Fifty-nine vertebrae from 20 lumbar spine specimens (mean age, 73 years ± 13; 11 women) were studied. FEA-estimated FL correlated with BMD (r = 0.75, p <.001) and was significantly lower in osteoporotic/osteopenic vertebrae (1222 ± 566 vs. 2880 ± 1182, p <.001). Dark-field and attenuation signals were positively correlated with FEA-estimated FL, in both vertical (r_darkfield = 0.64, p <.001, r_attenuation = 0.82, p <.001) and horizontal position (r_darkfield = 0.55, p <.001, r_attenuation = 0.81, p <.001). Conclusion: Dark-field and attenuation signals assessed using a clinical X-ray dark-field system significantly correlated with FEA-estimated FL in human spine specimens with and without osteoporosis/osteopenia. Dark-Field imaging may complement existing assessment methods for bone strength as a dose-efficient, accessible tool.