Background Bone mineral density (BMD) by dual-energy X-ray absortiometry (DXA) is still the routinely diagnostic approach for osteoporosis. However, BMD alone is not a good predictor of fracture risk while bone microarchitecture emerges as a determinant of bone fragility independent of BMD. High-resolution magnetic resonance imaging (MRI) represents an effective tool for in vivo characterization of trabecular bone microarchitecture (TBA) by noninvasive/nonionizing methods. Nevertheless, texture analysis is not used in clinical practice because of the large number of parameters to be calculated and analyzed.
We previously developed a MRI method to provide one parameter (TBLβ) sensitive to TBA changes in aging and osteoporosis1. Fractal lacunarity was chosen for TBA texture analysis as it is able to describe bone network discontinuity and sizes of bone marrow spaces, whose changes are index of increased fracture risk.
Objectives Clinical validation of the method as a new tool useful in early diagnosis of osteoporotic fracture risk is the strategic aim of LOTO study. One major objective is to verify the potential of TBLβ in discriminating patients with (VF+) and without (VF-) vertebral fractures. TBLβ as a potential marker of osteoporosis treatment efficacy is also investigated.
Methods An observational, cross-sectional and prospective study on over50s women at risk for bone fragility fractures was designed to validate the method2. Sample size was estimated equal to 280 osteopenic/osteoporotic women with/without prior vertebral bone fragility fractures. The main outcome measure is TBLβ as an index of osteoporotic fracture risk. It is calculated by a software prototype of the gray-scale version of our method on L4 axial images acquired by 1.5T MRI spin-echo multislice technique3.
Results A complete set of baseline recording, including DXA-BMD, conventional column Rx morphometry, and lumbar spine MRI-spin echo images for TBA characterization, was obtained for 279 out of 309 subjects eligible for the study. Prevalent VF were found in 31.5% subjects, 47.7% of which defined osteopenic at lumbar spine by DXA-BMD and 67% younger than 65 years. Baseline results from ROC analysis show that the contribution of TBA degeneration (TBLβ=40) to prevalent fractures is statistically higher (p=0.032) than BMD (T-score=-2.5). TBLβ results (Table 1) show that the proposed method is able to discriminate between VF+ and VF- patients (p=0.001). This result is further stressed in untreated T- subjects (p<0.0001). Treatment, any medication (T+), and drug therapy in particular, significantly counteract the difference between VF+ and VF- within groups (Table 1) and between groups with TBLβ values comparable to untreated VF- patients (p=0.319) and statistically higher than untreated VF+ (p=0.014).
Conclusions These promising results stress the usefulness of the method as a diagnostic tool in the assessment of osteoporotic fracture risk and suggest a potential role of TBLβ as a marker of treatment efficacy. More intriguing results are expected from prospective LOTO data.
Zaia et al. IEEE Trans Inf Technol Biomed 2006; 10:484.
Zaia et al. Osteoporos Int 2016; submitted.
Zaia. World J Orthop 2015; 6:221.
Disclosure of Interest None declared