|| Checking for direct PDF access through Ovid
Prevalent vertebral fractures confer a high risk of subsequent fractures; therefore, fracture identification is a key issue in osteoporosis management. Dual-energy X-ray absorptiometry has become the most widespread technique for the assessment of bone density, and can also provide images for assessment of vertebral morphometry (MXA) that are undistorted by parallax. The radiation dose is very low compared with conventional radiography. Whether the resolution of MXA is adequate for vertebral morphometry is uncertain. The present study compares MXA with conventional radiography in detecting vertebral deformities.MXA scans and lateral radiographs were performed in 61 women with osteoporosis. Adequate images were obtained in 99.1% of the 793 vertebrae by radiography and 95.6% by MXA. Poor image quality was mostly at T4–6 by MXA. Precision of repeat analyses was comparable between the techniques. Heights measured by MXA were 24 to 35% less than the radiographic values, the methods produced equivalent results for the ratio of anterior to posterior heights, but the MXA values for middle to posterior ratios were about 10% lower than for the radiographs (p < 0.0001). The extent of agreement between the techniques in identifying vertebral fractures was dependent on the strictness of the fracture definition used. With a 3 standard deviation (SD) criterion, the techniques found similar numbers of patients to have vertebral deformities, and the concordance in classification of individual vertebrae was 94%. At a 2 SD, cut-off concordance was 79%, and at 4 SD, it was 96%. Receiver-operating characteristic curves using cut-offs of ≥3 SDs also showed excellent diagnostic accuracy. We conclude that MXA shows acceptable performance for clinical use in diagnosing vertebral deformities, as long as cut-offs of ≥3 SDs are used, although a few percent of vertebrae in the upper thoracic region cannot be imaged adequately using this technique.