Olsen J B, Sager E M
Department of Medical Physics and Technology, Norwegian Radium Hospital, Montebello, Oslo.
Br J Radiol. 1995 Feb;68(806):160-4. doi: 10.1259/0007-1285-68-806-160.
A breast tissue phantom was used for assessment of mammographic image quality. Three images, exposed to give a reference area film density of 1.25 +/- 0.03, were obtained for each of five different film/kV combinations. Each of three radiologists carried out a series of three blind readings, in which the 15 films were ranked according to sharpness and contrast. In a similar experiment with a conventional image quality phantom, scores for the visualization of simulated calcifications and contrast-detail performance, and measures for resolution and contrast were obtained. In both experiments, the judgements of the readers were highly consistent, and both the intraobserver and interobserver variabilities were non-significant. No significant differences between the film/kV combinations were found with respect to high-contrast resolution and visualization of the simulated calcifications. The contrast measures obtained with the conventional phantom correlated well with the contrast scores obtained with the breast tissue phantom. The contrast-detail scores obtained with the conventional phantom showed no correlation with the scores obtained with the breast tissue phantom, nor with the contrast measures obtained with the conventional phantom.
使用乳腺组织模体评估乳腺摄影图像质量。对于五种不同的胶片/kV组合,每种组合都获取了三张曝光后参考区域胶片密度为1.25±0.03的图像。三位放射科医生各自进行了一系列三次盲法读片,将15张胶片按照锐利度和对比度进行排序。在使用传统图像质量模体的类似实验中,获取了模拟钙化可视化的分数、对比度细节性能以及分辨率和对比度的测量值。在两个实验中,读者的判断高度一致,观察者内和观察者间的变异性均无统计学意义。在高对比度分辨率和模拟钙化的可视化方面,未发现胶片/kV组合之间存在显著差异。使用传统模体获得的对比度测量值与使用乳腺组织模体获得的对比度分数具有良好的相关性。使用传统模体获得的对比度细节分数与使用乳腺组织模体获得的分数以及使用传统模体获得的对比度测量值均无相关性。