摘要
目的通比较CAD、医生双阅片、医生结合CAD阅片三种CT筛查肺结节方式,评价计算机辅助检测系统(CAD)在CT筛查肺结节中的应用价值。方法从CT扫描的无症状体检者中,应用随机抽样方法抽取60例肺结节患者。分别使用3种方式阅读图像,方法A:应用CAD软件对图像进行自动诊断;方法B:由2名放射科医生共同阅读图像并以一致意见作为诊断;方法C:由另外1名同等年资的放射科医生结合CAD软件阅读图像并做出诊断。结节参照标准由3位主任医生共同拟定。应用χ2检验分析三种阅读方式灵敏度是否存在差异,P<0.05为差异有统计学差异。结果 CAD、医生双阅片、医生结合CAD三种方式肺结节检出灵敏度分别为75.1%、73.8%、87.3%。医生结合CAD方式肺结节检出灵敏度显著高于CAD、医生双阅片方式(P=0.001,P=0.0003)。CAD与医生双阅片方式之间肺结节检出灵敏度没有显著差异(P=0.74)。结论 CAD系统可以提高放射科医生肺结节检出能力,且医生结合CAD诊断方式优于同年资医生双阅片方式。
Objective By comparing three reading methods of CT screening for lung nodules: CAD, double reading and the radiologist combined with CAD. To evaluation the application of computer aided testing system(CAD) in CT screening for lung nodules. Methods 60 subjects diagnosed with lung nodules undergoing the rountine-dose screening CT examination were random selected. Three methods were used, Method A: using CAD system to identify the image automatically. Methods B: two radiologists reading all images, record the results and with 2 agreement as the final diagnosis; Method C: by another radiologist as the same work experience with CAD system, recorded the detection as the final diagnosis. The reference standard was determined by three raiologists. χ2 test was performed to test whether there is significant difference between sensitivities of three reading modes, and statistical significances were defined as P〈0.05. Results CAD system, double reading, the radiologist with CAD method of pulmonary nodules detection sensitivity was 75.1%,73.8% and 87.3% respectively. The sensitivities of pulmonary nodule detection of the radiologist combined with CAD was significantly higher than CAD and double reading methods(P=0.001,P=0.0003).There were no significant sensitivity diffrernce between CAD and double reading method(P=0.74). Conclusion The CAD system can improve the ability of detection pulmonary nodules of the radiologist, and radiologists in combination with CAD diagnosis method is better than double reading method with same work experience.
出处
《中国CT和MRI杂志》
2016年第5期33-35,共3页
Chinese Journal of CT and MRI
关键词
X线计算机
肺结节
计算机辅助检测
X-ray Computed
Computer-aided Detection
Pulmonary Nodule