初步评估16层螺旋CT的Lung Care软件在肺结节研究中的辅助价值

Preliminary evaluation of Lung Care software of 16-slice helical CT in the study of pulmonary nodules

目的 探讨 1 6层螺旋CT的LungCare软件对肺结节的辅助诊断能力及适用范围。方法 低剂量CT扫描参数为 :1 2 0kV ,2 0emAs(effectivemAs ,有效毫安秒 ) ,床进 2 4或 36mm ,5 6例患者分为A、B 2组 ,采用准直 1 5mm ,重建间隔、重建层厚A组按 4、4mm ,B组按 2、2mm重建 ;1 2例患者为C组 ,采用准直 0 75mm ,重建间隔、重建层厚按 0 75、0 75mm重建图像序列。以A、B、C 3组横轴面图像发现的“肯定、模糊、难以判定”肺结节 ,再各分为 <5、5～ 1 0、>1 0～≤ 2 0mm组 ,用LungCare软件的旋转多平面重建 (r MPR)或感兴趣容积 (VOI)法三维显示功能观察结节是真的肺结节 ,还是肺血管断面等。将经三维显示确定的真的肺结节仍按 <5、5～ 1 0、>1 0～≤ 2 0mm组 ,用LungCare软件的快速测量肺结节容积及密度值分布功能进行测量 ,测量结果分为成功、失败。结果 LungCare软件分析聚集分布的多发肺结节困难 ;C组每个横轴面图像序列的图像过多 (达 381～ 4 72幅 ) ,阅读费时 ,不适于临床应用。LungCare软件的r MPR与VOI法三维显示模糊肺小结节存在一致性 ,但一致性不够理想 (Kappa =0 36 9,P =0 0 0 2 ) ;显示模糊肺结节 ,r MPR与VOI法的差异有统计学意义 ,r MPR法好于VOI法 (P =0 0 0 1 )。

Objective To evaluate the auxiliary diagnostic ability and applicability of the Lung Care software for the study of the pulmonary nodules.Methods Fifty-six patients underwent low-dose CT scan with 1.5 mm collimation, 4 mm reconstruction interval, and 4 mm reconstruction slice in group A, and with 1.5 mm collimation, 2 mm reconstruction interval, and 2 mm reconstruction slice in group B.12 patients underwent low-dose CT with 0.75 mm collimation, 0.75 mm reconstruction interval, and 0.75 mm reconstruction slice in group C.The nodules detected in groups A, B, and C were analyzed by r-MPR or VOI of the Lung Care software to distinguish the true pulmonary nodules from the vessels.The volume and density distribution of the true pulmonary nodules in groups A, B, and C were measured with the Lung Care software.Results It was difficult to observe the diffuse pulmonary nodules by r-MPR or VOI of the Lung Care software. The images of each patient in group C were too many to be applied in the clinic.There was statistically consistent in the observation of pulmonary nodules between r-MPR and VOI, but the coincidence was not good (Kappa=0.369, P=0.002).There was statistically significant difference in showing faint nodules between r-MPR and VOI (P=0.001), r-MPR was better than VOI.There was statistically significant difference between group A and B in showing <5 mm nodules by r-MPR and VOI of the Lung Care software (χ2=3.886, P=0.045), but no statistically significant difference in showing 5-10 mm nodules (χ2=0.170, P=0.680).The volume and density distribution of most 5-≤20 mm nodules were successfully measured with the Lung Care software, whereas those of most <5 mm nodules were not.There were statistically significant differences between group A and B in measuring <5 mm nodules (χ2=5.811, P=0.016) and 5-10 mm nodules (χ2=13.500, P<0.001) by the Lung Care software, but no statistically significant differences in measuring >10-≤20 mm nodules (χ2=0.000, P=1.000). Conclusion For distinguishing the true pulmonary nodules from others, the Lung Care software is suitable for the well-edged pulmonary nodules and most faint nodules, but not suitable for the nodules such as ground-glass opacity.For measuring the volume and density of the nodules, the Lung Care software is appropriate for the clear pulmonary nodules≥5 mm, and not appropriate for the faint or <5 mm nodules.The Lung Care software is not provided with automated lung nodule detection.