Comparison of two methods for evaluation of the image quality of lumbar spine radiographs

Two methods for visual evaluation of image quality of clinical radiographs have been compared. In visual grading analysis (VGA) specified anatomical structures in an image are visually compared with the same structures in a reference image, and in a free-response forced error (FFE) experiment - an extension of conventional ROC (receiver operating characteristics) analysis - the objective is to correctly localise known lesions. The spatial resolution and noise of digitised clinical radiographs of the lumbar spine were altered by image processing, and pathological structures were added to the images for the FFE experiment. The images were printed to film and evaluated by seven European expert radiologists using VGA and FFE. The results of these two different methods showed a very good agreement. In conclusion, VGA methodology can be made as solid as the FFE experiment for evaluating image quality. The simplicity of VGA makes it very suitable for implementation in clinical practice