Accuracy of bedside chest hard-copy screen-film versus hard- and soft-copy computed radiographs in a medical intensive care unit: receiver operating characteristic analysis.

Abstract

To compare the clinical diagnostic accuracy of hard-copy readings of screen-film bedside chest radiographs and both hard- and soft-copy readings of bedside chest computed radiographs obtained in a medical intensive care unit. Two samples of 95 cases were assembled from chest images obtained in 541 patients with either screen-film radiography or computed radiography. The cases were stratified according to the clinical problem for which the examination was ordered; the corresponding diagnosis was verified by a panel of two or three radiologists. Four radiologists blindly read the hard-copy images obtained with screen-film or computed radiography. Six months later, the radiologists read the computed radiographs by using an 8-bit, 1,684 x 2,048-pixel display. The data were analyzed by using multireader-multicase receiver operating characteristic (ROC) analysis of variance. No statistically significant differences in the area under the ROC curve were found between any of the methods. The results provide some justification for using bedside chest computed radiography and for reading soft-copy images from a high-quality display.