Measurement of vertebral body heights: Ex vivo comparisons between morphometric X-ray absorptiometry, morphometric radiography and direct measurements

Morphometric X-ray absorptiometry (MXA) offers some potential advantages ov er spinal radiography for the quantitative evaluation of vertebral fracture s in individuals with spinal osteoporosis. This ex vivo study examined the accuracy of MXA and quantitative morphometry (QM) for the evaluation of ver tebral height, in relation to direct measurements from cadaveric vertebral columns. Spinal radiographs and MXA scans were obtained from nine cadaveric vertebral columns (mean age at death 64 years). Anterior, middle and poste rior vertebral body heights of all segments from T4 to LA were measured int eractively using MXA software and QM (from the spinal radiographs), and com pared with direct measurements derived using digital callipers following ca daveric dissection. Coefficients of variation for repeat QM and MXA scan an alysis were less than 2%. The QM and MXA measurements were both strongly co rrelated (r(2) = 0.99) with the direct measurements, with accuracy errors o f between 2.6% and 4.3%, and 4.5% and 4.8%, for QM and MXA respectively. At the three measurement sites, analysis of measurement differences showed th at QM tended to overestimate the true height and MXA tended towards under-e stimation. The mean difference between the direct measurements and QM measu rements ranged from 2.7% to 8.7%, and the mean difference between direct an d MXA measurements ranged from -2.9% to 4.1%. There was strong linear assoc iation between the MXA and QM measurements (r(2) = 0.99) with mean differen ces at the three measurement sites ranging from -4.1% to -5.9%. Under the o ptimal (ex vivo) scanning conditions used in this study, MXA is comparable to spinal radiography for the assessment of vertebral height. Further studi es are required to determine the conditions under which satisfactory MXA im age resolution and measurement accuracy can be achieved in a clinical popul ation.