Rationale and objectives: The authors' purpose was to investigate the reliability of a personal computer (PC)-based display system compared with a workstation in the evaluation of rheumatoid arthritis on computed radiographs of the hands.
Materials and methods: Two radiologists on two occasions independently scored randomized computed radiographs of individual joints of the hands from 23 patients with rheumatoid arthritis and 14 control subjects. Each joint was scored from 0 (definitely normal) to 30 (severe disease) for each of four variables: soft-tissue swelling, osteopenia, erosions, and joint space narrowing. The observations were replicated on a picture archiving and communication system workstation and a PC. Intraobserver and interobserver reliability were calculated, as was the difference in scores between the two systems. The null hypothesis was that there was no difference between the workstation and the PC.
Results: The intraobserver reliability for normal versus abnormal joints was 73% with the workstation and 79% with the PC. The intraobserver reliability for workstation versus PC was 83%. There was moderate interreader reliability for both platforms (average kappa statistic, 0.46 [workstation] vs 0.45 [PC]). Small differences in scores between platforms are probably due mostly to the ordinal nature of the scoring system.
Conclusion: For evaluating computed radiographs of the hands in early rheumatoid arthritis, a PC-based system provides results similar to those obtained with a workstation, at considerably reduced cost.