A Comparison between Liquid-Crystal Display Monitors and Cathode-Ray Tube Monitors: A Combined Assessment Study of Observer Performances by Using Storage Phosphor and Flat-Panel-Detector Radiography in Detecting Experimentally Induced Pulmonary Edema in

Abstract

To compare the observer performance 0f liquid crystal display (LCO) and cathode ray tube (CRT) monitors in detecting experimentally induced pulmonary edema in pigs by using soft-copy images of amorphous selenium-based flat-panel-detector radiography (OR) and storage phosphor computed radiography (CR). Oleic acid was injected intra-atrially into three pigs (weight, 20-25kg) at doses of 0.04, 0.05, and 0.06 ml/kg to induce pulmonary edema. Each set of CR, DR, and thin-section CT scans were obtained every 20-30 minutes from three pigs over 4-6hours. Thus, 37 sets (10 sets from pig-1, 11 sets from pig-2, and 16 sets from pig-3) of radiographs were obtained. Images were masked for identity. randomly sorted, and displayed on both live mega pixel (2048 x 2560 x 8 bits) LCD and CRT monitors. Eight radiologists rated each image for Ihe presence of ill-defined diffuse opacities and reticular-Iinear opacities in both lungs by using continuous rating scale of 0-100. A total of 4736 (37 sets 2 detector system 2 fields 2 lesion types 8 observers 2 monitor systems) observations were analyzed in terms of receiver operating characteristics. Average observer performance in detecting ill defined diffuse opacities, LCO and CRT monitors were not different significantly in both OR and CR images. Average performance in detecting reticular-linear opacities was significantly better with LCD than CRT. These differences were significant in evaluating DR images (AUC=.852±.038 on LCD; AUC=.785±.070 on CRT) but not significant in evaluating CR images (AUC=.795±.060 on LCD; AUC=.745±.070 on CRT) System on both LCO and CRT monitors (p=.042 on LCD; p=.044 on CRT). Moreover. with OR system. observer performance was better with LCD monitor than with CRT monitor (p=.013), whereas with CR system, observer performance was not different significantly on both monitors (p=.11 8). Overall the five-megapixel LCD monitor was equal or superior to CRT monitor of the same pixel size in detecting experimentally induced pulmonary edema. Moreover, the LCD monitor appears, to be more optimized for detecting pulmonary reticular-linear opacity, when interfaced with DR system rather than with CR system.