Objective:To evaluate the accuracy and the consistency of CT phantom nodule volumetry measurement with pulmonary CT intellective CAD software and Lung VCAR sotware. Methods:Spherical nodules 2. 5mm, 5. 0mm, 10. 0mm and 20. 0mm in diameter were placed within a commercially available lung phantom and scanned by a 64 slice CT scanner. Nodule volume (V) was measured by the formula V= (4/3) 7trs. Automated nodule volume measurements were performed by using the two above mentioned CAD softwares independently. Results:The volume of all 16 nodules within the phantom could be measured and analyzed by the two softwares. The average relative errors of nodules 2. 5mm,5. 0mm, 10. 0mm and 20mm in diameter were 21. 1% ,-4. 7% , -1. 6% and -2. 3% respectively with pulmonary CT intellective CAD soltware. 1 ne average relative errors ol nodules 2.5mm,5.0mm, 10.0mm and 20.0mm in diameter were 15.5% 3. 1 % ,2. 7% and 1.4% respectively with Lung VCAR soltware. There was no substantial difference in the volume estima tions between the two CAD softwares by Bland Altman analysis. Conclusion: Although the accuracy of volumetry measure ment with pulmonary CT intellective CAD software was lower than that of lung VCAR,but still there was high consistency between the two CAD softwares.%目的:评估肺部CT智能诊断报告系统软件和Lung VCAR软件在测量体模结节体积方面的准确性和一致性.方法:使用植入直径为2.5、5.0、10.0和20.0mm的孤立结节的商用体模,采用64层螺旋CT机进行体模扫描.结节标准体积采用公式V=4/3πr3计算.分别使用肺部CT智能诊断报告系统软件和Lung VCAR软件对体模扫描数据进行分析,测量体模结节的体积.结果:两种软件对肺结节体模中16个结节均能进行相应的分析和体积测量.肺部CT智能诊断报告系统软件测量直径2.5、5.0、10.0和20.0mm结节体积的平均相对误差分别为21.1%、-4.7%、-1.6%和-2.3%;Lung VCAR软件的平均相对误差分别为15.5%、3.1%、2.7%和 1.4%.通过Bland-Altman分析,显示两种软件对体模结节体积测量的相对误差具有较好的一致性.结论:肺部CT智能诊断报告系统软件虽然比Lung VCAR软件在体模结节体积测量上准确性稍低,但两者对于体模结节体积测量具有较好的一致性.
展开▼
