Kehler M, Lyttkens K, Andersson B, Hochbergs P, Lindberg C G, Medin J, Nordström A J, Sanfridsson J, Vojciechowski J
Department of Radiology, Central Hospital, Karlskrona, Sweden.
Acta Radiol. 1996 May;37(3 Pt 1):332-6. doi: 10.1177/02841851960371P170.
Chest radiographs from 3 digital systems--2 based on luminescent phosphors and one on selenium--and a conventional film-screen system were evaluated and compared.
Computed radiography (CR) has for the past years been dominated by a single manufacturer, but now several systems have been marketed. Using a chest phantom and, as test objects, 2 simulated tumours for the lung and mediastinum, respectively, and one object simulating pulmonary lines, a total of 400 exposures were made, 100 on each system. The test objects were placed randomly with a ratio of presence/absence of each object of about 50. Six radiologists participated, 2 residents and 4 staff members. A receiver operating characteristics (ROC) analysis was performed with construction of curves, and the difference between the curves was estimated with a 2-tailed paired t-test.
The selenium-based system performed significantly better for pulmonary line detection than all the other systems, and better than one storage phosphor system for the lung "tumour" (p < 0.05), while one storage phosphor system was slightly better than the other in diagnosing all 3 test objects. The score for the film-screen system was only average.
