摘要
目的探讨人工智能迭代重组(AIIR)技术在病毒性肺炎不同分型中CT图像质量的应用价值。方法回顾性研究来我院就诊的新型冠状病毒性肺炎确诊患者86例,采用低剂量胸部CT扫描协议进行双肺CT扫描。分别对原始数据进行KARL 5级迭代重组、AIIR 3级重组,依据新冠肺炎诊疗指南第十版分为普通组、重症组,分别测量两组图像的正常肺组织、炎症区的CT值、图像噪声值(SD)、信噪比(SNR)。采用李克特(Likert)5分量表法对图像进行主观评价,应用统计学分析对图像质量进行客观评测。结果在同低剂量双肺CT扫描方案下,主观评分AIIR优于KARL重组。普通型肺炎组的正常肺组织区域、普通组和重症组的炎症区域的CT值在两种不同的重组方法下无统计学差异。AIIR后重症组炎症区域的SD值、SNR值高于KARL组(t=-11.33,P<0.05),余各组的SD值、SNR值均低于KARL组,且均有统计学差异(t=4.82,P<0.05)、(t=9.25,P<0.05)、(t=13.17,P<0.05)、(t=-2.05,P<0.05)。结论AIIR技术较KARL迭代重组技术可以实现较低的噪声以及较高的信噪比,提高图像的主观评分,有助于新型冠状病毒性肺炎的诊断。
Objective To explore the value of artificial intelligence iterative reconstruction(AIIR)technology in CT image quality of novel coronavirus(COVID-19)pneumonia.Methods According to the tenth edition of the COVID-19 diagnosis and treatment guidelines,86 patients with COVID-19 pneumonia were divided into ordinary and severe groups.Low-dose chest CT of all patients was reconstructed using KARL 5 iterative reconstruction and AIIR 3 technology.The CT values,standard deviations(SD)and signal-to-noise ratios(SNR)of normal and inflammatory lung areas were measured.Image quality was scored qualitatively using the 5-point Likert scale and analyzed statistically.Results The subjective score of AIIR was superior to KARL reconstruction.There was no significant difference between the two reconstruction methods in CT values of the normal lung in the ordinary pneumonia group,the inflammatory areas in the ordinary and severe pneumonia groups.The SD and SNR values with AIIR were significantly higher(t=-11.33,P<0.05)than those with KARL in the inflammatory areas of severe pneumonia and significantly lower in normal or inflamed lung of ordinary pneumonia(P<0.05).Conclusion AIIR is superior to KARL technology with less noise and higher SNR for CT diagnosis of COVID-19 pneumonia.
作者
马钦
龚欢
郭瑞
彭睿
宋法亮
马静
MA Qin;GONG Huan;GUO Rui;PENG Rui;SONG Faliang;MA Jing(Department of Radiology,Xinjiang Production and Construction Corps Hospital,Xinjiang 830000,China)
出处
《影像诊断与介入放射学》
2023年第5期342-346,共5页
Diagnostic Imaging & Interventional Radiology
基金
新疆生产建设兵团医院院级科技计划项目基金(2022002)。