Clinical diagnosis from digital displays: results and conclusions from the St Mary's evaluation project.

Abstract

Preliminary results with the first commercially available digital display system to be installed in a British radiology department were published in 1989: these consisted of data from observer performance studies of digitized images displayed using a 1024-line monitor, showing a single pathological feature--subperiosteal resorption in renal osteodystrophy. Further experiments have now been conducted with the successor to this equipment, a 1280-line digital display system. Formal observer performance studies were undertaken for four pathological conditions, and the results show statistically significant differences in performance between the digitized displayed images and those on film. The display system was not considered to be good enough for the task of primary radiological diagnosis of subtle lesions; findings support the conclusion that careful, objective clinical evaluation of digital display systems is important before they are introduced into clinical use.