Comparative Study of Dual-energy X‑ray Absorptiometry versus Multislice Computed Tomography for Vertebral Height Measurement

This study investigates the discrepancies between two-dimensional (2D) and three-dimensional (3D) imaging-based approaches for vertebral height assessment.

Lumbar spine CT scans and dual-energy X‑ray absorptiometry (DXA) studies within 6 months were reviewed. Vertebral fracture assessment was performed by converting CT-measured height ratios into Genant grades using three methods: conventional (A), consensus (B), and global search (C). Dataset 1 included 34 patients (191 vertebrae; 1,302 CT slices) with CT and DXA; Dataset 2 included 114 patients with CT only.

No significant difference was observed between Method A and DXA-based grading (p = 0.33). In contrast, Methods B and C significantly differed from DXA (p < 0.001, both). Discrepancies were greatest for vertebrae graded as 1 by Method A (98 and 166 vertebrae reclassified by Methods B and C, respectively). Compared with Method A, Method B revealed greater height reductions in Grades 0, 1, and 2 vertebrae by 9.00, 9.13, and 13.75 percentage points, respectively. For Method C, these reductions were 10.71, 10.14, and 12.35 percentage points, respectively.

Compared with DXA, continuous sagittal CT images improved detection of greatest vertebral height loss, especially for Grade 1 reductions. When applying 2D criteria to CT scans, clinicians should be cautious of a tendency to overestimate vertebral height loss.

This first study demonstrates that the 3D method yields vertebral height reduction 10.84 percentage points greater than the 2D method. Therefore, overestimation of vertebral height reduction on CT should be considered when radiological classification criteria are applied.