Objective: Evaluate the characteristics of 3D DXA-parameters at the lumbar vertebrae in patients whose had suffered vertebral fracture comparing to non-fractured subjects. Instead evaluate differences between thoracic and lumbar fractures.
Method: We retrospectively analyzed 28 patients – age range (37.1 -81.1 years old) and fracture group ( 37.05-86.1years old) : Thoracic: 15 only thoracic fractures; Lumbar: 13 at least one lumbar fracture. aBMD at L1-L4 segment was measured by StratosDr ( DMS Apelem ) . It has integrated 3D reconstruction software for the lumbar spine (version V4.1). The software, developed by Galgo, estimates the vBMD and 3D shape of the lumbar spine from an AP DXA image by registering a statistical model onto the 2D-DXA image The analyzed variables: global volumetric bone density throughout the vertebra [GvBMD], volumetric BMD of the integrated bone [IntvBMD], trabecular bone [TvBMD], cortical bone [CvBMD] and average thickness of cortical shell, in the vertebrae L1 to L4. Statistical analysis: T-Test for mean values and the sensitivity / specificity of the measurements by means of Receiver Operating Characteristic (AUC).
Results: All parameters evaluated were significant lower at fractured group compared to non-fractured subjects. 3D parameters measured at vertebral body were trabecular vBMD, -14.3%, (113,3 vs 132,6 g/cm3 p< 0,07 )while cortical component showed only -6% (539.9 vs 574.5mg/cm3; p<0.003). At the thoracic fractures fractured patients had -20% less TvBMD (106.8 vs 132.6mg/cm3p<0.004) and 8% less at CvBMD (528.5 vs 574.5mg/cm3; p< 0.002). While in lumbar spine fractures differences at TvBMD was 9.09% (120.8 vs 132.6 mg/cm3; p: NS), however at CvBMD the difference was lower (3.6%) but significant (553.1 vs 574.5mg/cm3).
Conclusion: DXA- based 3D modeling parameters are associated in a way similar to classic 2D measurements. The association of trabecular component are of greater magnitude at Thoracic Fractures, while GvBMD is most compromise at Lumbar Fractures.