肺磨玻璃结节多层CT自动体积和密度测量的可重复性研究

Pulmonary ground-glass nodules:In vivo repeatability of automated volume and density calculations with MDCT

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-Altman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。

Objective The objective of our study was to evaluate the in vivo reproducibility of automated volume and den- sity calculations of pulmonary ground-glass nodules with MDCT. Methods Thirty-five subjects with 46 persistent pulmo- nary 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 availa- ble 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 950//00 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 vol- ume or in density.