Background Fracture risk prediction in patient relies chiefly on bone mineral density (BMD) measurement. However, the decreased bone strength characteristic of osteoporosis is dependent not only on BMD, but also on other factors, most notably bone microarchitecture.

Objectives The purpose of this study was to investigate bone microarchitecture variables of cadaveric vertebrae using ultra-high field MRI (7 Tesla).

Methods Twenty four vertebrae (L2, L3, L4) from eight cadavers were studied using 7 Tesla MRI. Their Bone Mineral Density (BMD) were investigated using dual energy X Ray absorptiometry. Then, all specimens underwent mechanical compression tests to failure and the failure load (in Newton) and constraint (in Mpa) were measured. Bone Volume Fraction (BV/TV), Trabecular Thickness (Tb.Th), and Trabecular Spacing (Tb.Sp) were measured in MR images using a Digital topological analysis (Bone J). Measurements were performed by two observators in order to characterize the inter-rater reliability. Statistical analyses were performed using SPSS. Correlations between variables were analyzed using Spearman correlations and Stepwise regression. A p value of 0.05 was considered as significant.

Results The inter-rater reliability for bone microarchitecture parameters quantification was good. Tb.Th and Tb.Sp measured using high-field MRI were 0.52±0.18 and 0.48±0.10 respectively while the BV/TV fraction was 0.52±0.13. The mean BMD was 0.86±0.20 g/cm². The failure load and the constraint measured during the compression tests were 2600±1267N and 1.57±0.81 Mpa respectively. Interestingly, the variables measured during the mechanical tests were significantly

The failure load and constraint measured during the compression tests were significantly correlated with the BMD. Regarding the bone indices quantified using high-field MRI, a significant linear relationship was observed between the trabecular spacing and the BMD (R²=0.23, p=0.01 and the constraint values to failure (R²=0.18, p=0.04). A stepwise regression with backward elimination demonstrated that combining BV/TV and BMD improved the relationship with the constraints from an adjusted R²=0.384 for BMD alone to an adjusted R²=0.41 for BMD + BV/TV.

Conclusions In the present study, we demonstrated for the first time that the variables characterizing the vertebral bone microarchitecture quantified using ultra-high field MRI were significantly correlated with biomechanical parameters. In addition, we illustrated that the vertebral bone strength was better described by a variable combining BMD and trabecular bone spacing.

Disclosure of Interest None declared