摘要
目的:探讨人工智能CT定量分析技术在肺挫伤病情进展中的应用价值。方法:回顾性分析90例肺挫伤患者108处病灶的CT资料,利用人工智能CT定量分析技术自动分割图像,并测量不同时期(急性期、渗出期及吸收好转期)肺挫伤3~7级支气管定量指标,行统计学分析。结果:渗出期3~7级支气管的最大壁厚、平均壁厚大于急性期和吸收好转期,渗出期4~7级支气管的管腔内外直径、管腔内外周长及5~7级支气管的内腔横截面积、管壁横截面积小于急性期和吸收好转期(均P<0.05)。急性期4~7级支气管的最大壁厚、平均壁厚大于吸收好转期而小于渗出期,急性期4~7级支气管的管腔内外直径、管腔内外周长及5~7级支气管的内腔横截面积、管壁横截面积小于吸收好转期而大于渗出期,差异均有统计学意义(均P<0.05)。急性期肺挫伤灶占两肺总容积的百分比为(4.26±1.81)%,渗出期为(7.91±2.70)%,吸收好转期为(1.63±0.91)%。急性期肺挫伤灶容积占比与7级支气管内腔横截面积、最大壁厚及平均壁厚相关,渗出期肺挫伤灶容积占比与5级支气管内腔横截面积、支气管内外直径及内周长相关,吸收好转期肺挫伤灶容积占比与6级支气管内腔横截面积、最大壁厚、平均壁厚及7级支气管最大壁厚相关。结论:人工智能CT定量分析技术可直观显示肺挫伤及支气管的形态学变化情况,为临床预估病情发展及干预治疗提供新方法。
Objective:To evaluate the potential of artificial intelligence CT quantitative analysis technology in tracking the progression and healing phases of pulmonary contusions.Methods:The chest CT data of 90 patients with pulmonary contusions(108 lesions)were retrospectively analyzed.Artificial intelligence CT quantitative analysis technology was used to automatically segment the images and measure the changes in the quantitative indicators of grade 3~7 bronchi in different periods of pulmonary contusion(acute stage,exudation stage and absorption stage),and the results were statistically analyzed.Results:Compared with the acute stage and absorption stage,the maximum and average wall thicknesses of grade 3~7 bronchi in the exudation stage were greater,while the intra-and extraluminal diameters and circumferences of grade 4~7 bronchi and the cross-sectional areas of luminal and wall of garde 5~7 bronchi in the exudation stage were significantly smaller(all P<0.05).The maximum and average wall thicknesses of grade 4~7 bronchi in the acute stage were significantly greater than those in the absorption stage but significantly smaller than those in the exudation stage,while the intra-and extraluminal diameters and circumferences of grade 4~7 bronchi and the cross-sectional areas of the luminal and wall of grade 5~7 bronchi in the acute stage were significantly smaller than those in the absorption stage but significantly larger than those in the exudation stage(all P<0.05).The volume proportion of pulmonary contusion was(4.26±1.81)%in the acute stage,(7.91±2.70)%in the exudation stage and(1.63±0.91)%in the absorption stage.The volume proportion of pulmonary contusion in the acute stage was correlated with the luminal cross-sectional area,the maximum and average wall thickness of grade 7 bronchi.The volume proportion of pulmonary contusion in the exudation stage was correlated with the luminal cross-sectional area,the intra-and extraluminal diameters,the intraluminal circumference of grade 5 bronchi.The volume proportion of pulmonary contusion in the absorption stage was correlated with the luminal cross-sectional area,the maximum and average wall thickness of grade 6 bronchi and the maximum wall thickness of grade 7 bronchi.Conclusions:Artificial intelligence CT quantitative analysis can be used to visualize bronchial morphological changes in various stages of pulmonary contusions and can be used to dynamically monitor the treatment of pulmonary contusions,providing a new approach for evaluating clinical efficacy.
作者
代永亮
张宇哲
张元刚
苟杰
程永涛
段小艺
郭佑民
李艳
DAI Yongliang;ZHANG Yuzhe;ZHANG Yuangang;GOU Jie;CHENG Yongtao;DUAN Xiaoyi;GUO Youmin;LI Yan(Department of Imaging,Hospital of Xidian University,Xi’an 710000,China;Department of Imaging,521th Hospital of Ordnance Group,Xi’an 710065,China;Department of Medical Imaging,First Affiliated Hospital of Xi’an Jiaotong University,Xi’an 710061,China)
出处
《中国中西医结合影像学杂志》
2024年第2期184-187,197,共5页
Chinese Imaging Journal of Integrated Traditional and Western Medicine
基金
陕西省科技厅重点研发计划项目(2022SF-344)。
关键词
肺挫伤
人工智能
体层摄影术
X线计算机
定量分析技术
Pulmonary contusion
Artificial intelligence
Tomography,X-ray computed
Quantitative analysis technology