Iwano S, Makino N, Ikeda M, Itoh S, Ishihara S, Tadokoro M, Ishigaki T
Department of Radiology, Toyota Memorial Hospital, Aichi, Japan.
J Comput Assist Tomogr. 2000 Mar-Apr;24(2):242-6. doi: 10.1097/00004728-200003000-00010.
The goal of this work was to determine a radiologist's ability to detect solitary pulmonary nodules on helical CT using both video (cine) viewing and film-based viewing.
Sixty-five chest helical CT studies were reviewed. Six radiologists searched for 40 lung nodules on CT images presented in three formats. Film-based viewing of images at 10 and 5 mm increments was performed with a light box. Video viewing of the same examinations was performed in 5 mm increments at 2 frames/s. The area under the receiver operating characteristic curve (Az) measured the observer's ability to detect nodules.
The Az was 0.948 for the video viewing, 0.844 for 5 mm increment film-based viewing, and 0.879 for 10 mm increment film-based viewing. There were no statistically significant differences.
Lung nodules can be detected with similar detection rates when viewing conventional film or videotaped helical CT images. Videotaped images incur a lower cost, an important consideration in mass screening for lung cancer.
本研究旨在确定放射科医生通过视频（电影）观看和基于胶片观看两种方式在螺旋CT上检测孤立性肺结节的能力。
回顾了65例胸部螺旋CT研究。六位放射科医生在以三种格式呈现的CT图像上寻找40个肺结节。使用灯箱以10毫米和5毫米的增量对图像进行基于胶片的观看。以每秒2帧的速度以5毫米的增量对相同检查进行视频观看。接收者操作特征曲线下面积（Az）衡量观察者检测结节的能力。
视频观看的Az为0.948，5毫米增量基于胶片观看的Az为0.844，10毫米增量基于胶片观看的Az为0.879。无统计学显著差异。
观看传统胶片或录像螺旋CT图像时，肺结节的检测率相似。录像图像成本较低，这在肺癌大规模筛查中是一个重要考虑因素。
