| Popis: |
Background A new method for the measurement of bone density of the phalanx using digitised conventional X-ray films of the hand has been developed in our centre. Objectives To test the reliability and internal validity (test-retest) of this method, especially regarding the digitalisation process and the software developed for segmentation and optical density analyses. Methods The non-dominant hand is radiographed together with an aluminium wedge used for calibration and correction of the variability associated with the XR technique and digitalisation process. A single exposure at 46 kV is performed and the resulting film is read with an AgfaScan T1200 device with predefined settings. Results To assess the reproducibility of our technique, XR hand films of 17 patients were scanned 5 times in different days. CV was found to be 0,913 (95% CI 0,650–1,825) for the middle phalanx; 0,632 (95% CI 0,470–1,265) for the proximal phalanx; and 1,194 (95% CI 0,918–2,387) for the central area of the third metacarpal. A test-retest experiment with 2 films in 50 patients gave an almost perfect internal correlation (Pearson’s R of 0,985; 0,995 and 0,986) for these same sites. We also validated our technique in a sample of 171 women referred by general practitioners with a clinical diagnosis of osteoporosis. Construct validation was performed against conventional lumbar spine and hip DXA (Lunar DPX). Also,78 of these women had a phalanx DXA measurement with a commercial device (AccuDXA). Correlations between BMD in several sites are described below: View this table: Abstract SAT0203 Table 1 Conclusion The reproducibility and internal correlation of measurements made with our newly developed technique is very high for the three areas of the hand studied. Correlation of BMD of the hand with BMD in the lumbar spine and hip is moderate and similar to that found among these places themselves. |
