Reproducibility and accuracy in the morphometric and mechanical quantification of trabecular bone from 3 Tesla magnetic resonance images

Abstract Objective We used an animal model to analyze the reproducibility and accuracy of certain biomarkers of bone image quality in comparison to a gold standard of computed microtomography (μCT). Materials and methods We used magnetic resonance (MR) imaging and μCT to study the metaphyses of 5 sheep tibiae. The MR images (3 T) were acquired with a T1-weighted gradient echo sequence and an isotropic spatial resolution of 180 μm. The μCT images were acquired using a scanner with a spatial resolution of 7.5 μm isotropic voxels. In the preparation of the images, we applied equalization, interpolation, and thresholding algorithms. In the quantitative analysis, we calculated the percentage of bone volume (BV/TV), the trabecular thickness (Tb.Th), the trabecular separation (Tb.Sp), the trabecular index (Tb.N), the 2D fractal dimension (D 2D ), the 3D fractal dimension (D 3D ), and the elastic module in the three spatial directions ( Ex , Ey and Ez ). Results The morphometric and mechanical quantification of trabecular bone by MR was very reproducible, with percentages of variation below 9% for all the parameters. Its accuracy compared to the gold standard (μCT) was high, with errors less than 15% for BV/TV, D 2D , D 3D , and E app x , E app y and E app z . Conclusions Our experimental results in animals confirm that the parameters of BV/TV, D 2D , D 3D , and E app x , E app y and E app z obtained by MR have excellent reproducibility and accuracy and can be used as imaging biomarkers for the quality of trabecular bone.