The impact of technical conditions of X-ray imaging on reproducibility and precision of digital computer-assisted X-ray radiogrammetry (DXR).

Abstract

To evaluate the reproducibility of imaging and analysis for bone mineral density (BMD) determination using digital computer-assisted X-ray radiogrammetry (DXR; Pronosco X-posure, version V.2, Sectra Pronosco, Denmark); to verify potential factors that influence BMD extrapolation such as tube voltage, film-focus distance (FFD), film quality and brand (Kodak T-MAT-Plus, Konika SRH, Agfa Scopix), imaging technology (conventional, digital), imaging system (Kodak, Agfa) and exposure level (mAs); and to clarify whether DXR analysis based on printouts of digital images is comparable to analysis of conventional images. The hand of a cadaver was X-rayed using varied parameters: 4-8 mAs, 40-52 kV, 90-130 cm FFD. Radiographs under standardised conditions were performed 10 times using a conventional machine (Philips Super 80 CP) and the printouts of a digital system (Digital Diagnost Philips Optimus) for the analysis of reproducibility. One image was scanned and analysed 10 times additionally for imaging reproducibility. Reliability error of the system for the imaging process using conventional radiographs-rays was 0.49% (standard conditions: 6 mAs, 40 kV, 1 m FFD), using printouts of digital images was 2.89% (4 mAs, 42 kV, 1 m FFD) and regarding the analysis process was 0.22%. BMD calculation is not affected by alterations in FFD (precision error 1.21%), mAs (0.83%) or film quality/brand (0.38%), but differs significantly depending on tube voltage (2.70%). The system was not able to analyse conventional images with tube voltages of 49/52 kV. DXR technology is stable with most of the tested parameters. Normative data should exclusively be used for calculations using similar tube voltage or correction factors. All other parameters had no significant influence on the BMD calculation. Reproducibility is high. For technical reasons it is not recommended to use the printouts of digital images for BMD determination.