Anthropomorphic chest phantom imaging – The potential for dose creep in computed radiography

Abstract For film-based radiography the operator had to be exact in the selection of acquisition parameters or the image could easily become under- or over-exposed. By contrast, digital technology allows for a much greater tolerance of acquisition factor selection which would still give an image of acceptable diagnostic quality. In turn this greater tolerance allows for the operator to increase effective dose for little or no penalty in image quality. The purpose of this article is to determine how image quality and lesion visibility vary with effective dose (E) in order to identify how much overexposure could be tolerated within the radiograph. Using an anthropomorphic chest phantom with ground glass lesions we determined how perceptual image quality and E varied over a wide range of acquisition conditions. Perceptual image quality comprised of image quality and lesion visibility. E was calculated using Monte Carlo method; image quality was determined using a two alternative forced choice (2AFC) method and the quality criteria were partly informed from European guidelines. Five clinicians with significant experience in image reading scored the images for quality (intraclass correlation coefficient 0.869). Image quality and lesion visibility had a close correlation ( R 2  > 0.8). The tolerance for over-exposure, whilst still acquiring an image of acceptable quality, increases with decreasing kV and increasing source to image distance (SID). The maximum over-exposure factor (ratio of maximum E to minimum E that produce images of acceptable quality) possible was 139 (at 125 cm and 60 kV). Given the phantom had characteristics similar to the human thorax we propose that that potential for overexposure in a human whilst still obtaining an image of acceptable perceptual image quality is very high. Further research into overexposure tolerance and dose creep should be undertaken.