[Computed radiography in X-ray examination of stomach: evaluation by phantom experiments and clinical studies]

The usefulness and clinical applicability of computed radiography (FCR) in X-ray examination of the stomach were evaluated in comparison with film-screen (FS) radiography by both phantom experiments and clinical studies. Receiver operating characteristic (ROC) curve analysis consisting of 25 normal and 25 artificially created simulated lesions on an originally created Styrofoam phantom showed no difference in lesion detectability between FCR and FS radiography. Both were also equal in evaluation of the internal structure and margin of the 25 lesions by visual ranking. ROC curve analysis of clinical cases was performed in 30 cases, 20 normal and 10 abnormal, while evaluation by visual ranking was performed on 18 lesions. There was no significant difference between FCR and FS radiography either with ROC curve analysis or visual ranking, although the variations were somewhat greater in the clinical cases than in the phantom studies. I conclude that FCR is equal to FS radiography in X-ray examination of the stomach regarding both the detectability and visual conspicuity of the lesion. In view of the other advantages of FCR, it can safely replace conventional FS radiography in X-ray examination of the stomach.