Dual-energy estimates of volumetric bone mineral densities in the lumbar spine using quantitative computed tomography better correlate with fracture properties when compared to single-energy BMD outcomes

Abstract It is estimated that over 200 million people worldwide are affected by osteoporosis. Vertebral fracture risk prediction using dual energy x-ray absorptiometry (DXA) is confounded by limitations of the technology, such as 2D measurements of bone mineral density (BMD), inability to measure bone distribution and heterogeneity, and potential overestimations of BMD due to degenerative diseases. To overcome these shortcomings, single energy (SE) quantitative computed tomography (QCT) imaging estimates of Hounsfield units (HU) and volumetric BMD have been implemented as alternative methodologies for assessing fracture risk. However, marrow fat within the vertebrae can highly affect the vBMD and fracture properties estimations. To address this issue, 54 vertebrae were dissected from nine cadaveric spines and scanned using SE-QCT (120kVp) and dual energy (DE)-QCT (80/140 kVp), with the latter accounting for marrow fat within the vertebrae. The vertebrae were then scanned using DXA and subjected to mechanical testing to obtain fracture properties. aBMD outcomes from DXA showed a better correlation with DE-QCT vBMD versus SE outcomes [DE: aBMD vs. vBMD (R2: 0.61); SE: aBMD vs. vBMD (R2: 0.27)]. SE-QCT underestimated vertebral vBMD by -56%; p