Objective The objective of our study was to evaluate the in vivo reproducibility of automated volume and density calculations of pulmonary ground-glass nodules with MDCT. Methods Thirty-five subjects with 46 persistent pulmonary ground-glass nodules (GGN) between 5 and 20 mm in diameter were enrolled in this prospective study. Two same MDCT data sets were obtained for each nodule on separate breath-holds during the same session. The repeated scan only covered the GGNs instead of the whole lung. The volume and density of each nodule were calculated by commercial available automated software. Repeatability was evaluated only in well segmented cases by Bland-Altman's approach. Results Automated volume and density calculations were succeeded with good segmentation in 78. 3% (36/46) nodules with both the two MDCT data sets. The 95% limits of agreement were 0. 984(0. 802. 1. 167)for nodule volume and 0. 996(0. 946. 1. 046)for nodule density. Conclusion A volume variation of greater than 27% and a density variation of greater than 7% for pulmonary GGNs between 5 and 20 mm confirmed by follow-up CT can be sure that a nodule is actually growing in volume or in density.%目的 利用多层CT对肺磨玻璃结节(ground-glass nodule,GGN)自动体积、密度测量的可重复性进行研究,为该技术的进一步临床应用提供参考.方法 对35名患者共计46个直径5～20mm持续存在的GGN进行重复2次多层CT扫描,第二次扫描范围仅覆盖结节附近区域.利用商用软件自动计算GGN体积和密度,选择对测量结果进行一致性分析.统计学方法采用Bland-Alrman方法.结果 两组数据GGN自动体积和密度测量均成功且结节分割比较完美的占78.3％ (36/46),体积测量值95％一致性界限为0.984(0.802,1.167),密度测量值95％一致性界限为0.996(0.946,1.046).结论 对于大多数GGN自动体积、密度测量的可重复性很好,体积增加27％、密度增加7％可以确认GGN的生长.
展开▼
