Suppr超能文献

[计算机断层扫描中肝脏病变识别的ROC(受试者工作特征)研究]

[ROC (receiver operating characteristic) study for the recognition of liver lesions on the computer tomogram].

作者信息

Rockstroh G, Rotte K H, Kriedemann E, Cobet H, Boitz F, Cimanowski N

出版信息

Digitale Bilddiagn. 1985 Jun;5(2):70-3.

PMID:4017439
Abstract

ROC examination was based on patient computer tomograms, in which lesions of the liver were simulated by image manipulation. The window width for optimal recognition of the lesions was found to be 128 Hounsfield units (HU). Evaluation of Receiver Operating Characteristics (ROC) by diameters and contrasts of the lesions yielded a contrast detail diagram largely reflecting conditions in clinical practice. The detectability--measured via the smallest visible diameter of the lesions--is inferior by the factor 3 to 60 to the results obtained with hole pattern phantoms. For example, in computer tomograms lesions of the liver of 24 mm and 4 HU or also 11 mm and 10 HU yield 70% true positive and 10% false positive findings.

摘要

ROC检查基于患者的计算机断层扫描,其中通过图像处理模拟肝脏病变。发现用于最佳识别病变的窗宽为128亨氏单位(HU)。通过病变的直径和对比度对受试者工作特征(ROC)进行评估,得出的对比细节图很大程度上反映了临床实际情况。通过病变的最小可见直径测量的可检测性比使用孔型体模获得的结果低3至60倍。例如,在计算机断层扫描中,24毫米和4 HU的肝脏病变或11毫米和10 HU的肝脏病变产生70%的真阳性和10%的假阳性结果。

文献检索

告别复杂PubMed语法,用中文像聊天一样搜索,搜遍4000万医学文献。AI智能推荐,让科研检索更轻松。

立即免费搜索

文件翻译

保留排版,准确专业,支持PDF/Word/PPT等文件格式,支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述,25分钟生成高质量综述,智能提取关键信息,辅助科研写作。

立即免费体验