CT temporal subtraction for detection of lung nodules

VernacularTitle:CT时间减影技术在肺结节检测中的应用
Author: Yunlong WANG ; Li FAN ; Yun WANG ; Yu GUAN ; Yi XIA ; Qiong LI ; Yi XIAO ; Shiyuan LIU
Keywords: lung tumor; image processing; computer-aided
From: Journal of Practical Radiology 2017;33(8):1276-1280
CountryChina
Language:Chinese
Abstract: Objective To evaluate the effect of CT temporal subtraction(TS) for detection of lung nodules.Methods 80 cases of CT images (current and previous CT images) and corresponding CT TS images were presented, which included 30 cases with nodules(75 nodules) and 50 controls.4 observers, including 2 radiologists and 2 residents, indicated their confidence level regarding the presence of a new emerging or larger or solid component than before,which first used standard CT images, then with the addition of CT TS images.Receiver operating characteristic(ROC) curve analysis was used to evaluate the observers'' performance.Results The mean value of the area under the ROC curve (AUC) were 0.860 and 0.925 for four observers without and with TS images,respectively(P<0.01).Average sensitivity for detection of lung nodules was improved from 77.3%(58/75) to 89.3%(67/75) by using CT TS images.Conclusion The use of CT TS images can significantly improve radiologists'' performance for detection of lung nodules,especially when a small size solid or subsolid nodule presents near the pulmonary hilum.CT TS is more helpful for residents.