CT肺结节分析软件在肺部筛查中的应用价值

目的:对比分析肺结节分析软件在肺部筛查中的应用价值。方法:80例肺部结节患者随机均分为A、B、C、D 4组。3位放射科医师(D1、D2、D3)分别采用常规阅片法和计算机辅助检测(computer aided detection,CAD)联合阅片法,分别在无限时和限时模式下,对这4组患者行CT阅片,记录肺部结节数量。其中A组采用无限时常规阅片法,B组采用无限时CAD联合阅片法,C组采用限时常规阅片法,D组采用限时CAD联合阅片法。将所有记录的肺结节结果与金标准进行对比,绘制受试者工作特性曲线(ROC曲线),记录曲线下面积(Az值)。通过Az值的比较,判断肺结节分析软件对于诊断者的辅助价值。结果:D1、D2、D3的Az值A组分别为0.777、0.693、0.506,B组分别为0.816、0.759、0.599,C组分别为0.666、0.559、0.295,D组分别为0.779、0.744、0.542。各诊断者于无限时模式下的Az值均高于限时模式下的Az值,差异具有统计学意义。在2种模式下,CAD联合阅片法的Az值均较常规阅片法升高,但是在无限时模式下差异无统计学意义,在限时模式下差异有统计学意义;Az值的升高有随着诊断经验的增加而下降的趋势。结论:肺结节分析软件能够帮助提高诊断者在肺结节检测中的诊断价值,尤其对快速诊断及低年资诊断者的帮助意义更大。

The application value of CT lung nodules analysis software in prompt diagnosis

Objective：To comparatively analyze the application value of lung nodules analysis software in timed reading. Methotis.. 80 thoracic CT examinations with pulmonary nodules were separated to 4 groups （A,B,C,D）. Three doctors of radiologists （D1 ,D2 ,D3） with different levels of experience read the four groups CT imaging and record the number of lung nodules found, with two methods of general reading or CAD reading,in two models of free time or time limited. The doctors analyzed group A with general reading method in free time model, group B with CAD reading method in free time model, group C with general reading method in time limited model and group D with CAD reading method in time limited model. The recorded results of lung nodules were compared with gold standard,drawing the ROC curve and calculating the Az average. By the comparison of Az av- erage,judged the value of lung nodules analysis software to diagnosticians. Results.. The Az average of diagnosticians （D1, D2, D3） were 0. 777,0. 693,0. 506 in group A,0. 826,0. 772,0. 620 in group ]3,0. 816,0. 759,0. 599 in group C,0. 779,0. 744,0. 542 in group D. The Az averages of diagnosticians in free time model were significantly higher than time limited model. In two models,the Az averages of CAD reading method were higher than general reading. But there were no significant differences in free time model,while significant differences were found in time limited model. The Az average rising of diagnostician had declining trend with the increasing of diagnosis experience. Conclusions： Lung nodules analysis software can help diagnostician to raise the diagnostic value in detecting pulmonary nodules, especially in rapid diagnosis. The lung care software is more helpful to low period service doctors.