Assessment of 2D and 3D fractal dimension measurements of trabecular bone from high-spatial resolution magnetic resonance images at 3 T

Purpose: In vivo two-dimensional (2D) fractal dimension (D-2D) analysis of the cancellous bone at 1.5 T has been related to bone structural complexity and shown to be a potential imaging-based biomarker for osteoporosis. The objectives of this study were to assess at 3 T the in vivo feasibility of three-dimensional (3D) bone fractal dimension (D-3D) analysis, analyze the relationship of D-2D and D-3D with osteoporosis, and investigate the relationship of D-3D with spinal bone mineral density (BMD). Methods: A total of 24 female subjects (67 +/- 7 yr old, mean +/- SD) was included in this study. The cohort consisted of 12 healthy volunteers and 12 patients with osteoporosis. MR image acquisitions were performed in the nondominant metaphysis of the distal radius with a 3 T MR scanner and an isotropic resolution of 180 mu m. After segmentation and structural reconstruction, 2D and 3D boxcounting algorithms were applied to calculate the fractal complexity of the cancellous bone. D-2D and D-3D values were compared between patients with osteoporosis and healthy subjects, and their relationship with radius BV/TV and spinal BMD was also assessed. Results: Significant differences between healthy subjects and patients with osteoporosis were obtained for D-3D (p < 0.001), with less differentiation for D-2D (p=0.04). The relationship between fractal dimension and BMD was not significant (r=0.43, p=0.16 and r=0.23, p=0.48, for D-2D and D-3D, respectively). Conclusions: The feasibility of trabecular bone D-3D calculations at 3 T and the relationship of both D-2D and D-3D parameters with osteoporosis were demonstrated, with a better differentiation for the 3D method. Furthermore, the D-3D parameter has probably a different nature of information regarding the trabecular bone status not directly explained by BMD alone. Future studies with subjects with osteopenia and larger sample sizes are warranted to further establish the potential of D-2D and D-3D in the study of osteoporosis. (C) 2010 American Association of Physicists in Medicine. [DOI: 10.1118/1.3481509]