能谱CT多参数成像改善新型冠状病毒肺炎图像质量的价值  

The value of multi-parameter imaging of spectral CT in improving image quality of Corona Virus Disease 2019

作　　者：李万江[1] 文雨婷 李真林[1] 帅桃[1] 廖凯[1] 羊丹 潘雪琳[1] 余建群[1] LI Wanjiang;WEN Yuting;LI Zhenlin;SHUAI Tao;LIAO Kai;YANG Dan;PAN Xuelin;YU Jianqun(Department of Radiology,West China Hospital,Sichuan University,Chengdu 610041,China)

机构地区：[1]四川大学华西医院放射科,四川成都610041

出　　处：《西部医学》2020年第8期1093-1097,共5页Medical Journal of West China

基　　金：四川省科技厅重点研发项目(2019YFS0522);四川大学华西医院学科卓越发展1-3-5工程项目(ZYGD18019)。

摘　　要：目的探讨能谱CT多参数成像改善新型冠状病毒肺炎(COVID-19)图像质量的价值。方法纳入2020年1月27日~2月20日确诊为COVID-19的15例患者胸部能谱CT影像学资料,共发现53处磨玻璃样病灶(磨玻璃结节组),23处实性病灶(实性结节组)。使用患者能谱CT的原始数据分别重建混合能量图像、最佳单能量图像和水基图像,在三种图像同一位置以相同大小(ROI)测量病灶的CT值(水基图测量水值)、相邻正常肺组织的CT值和标准差(SD),计算信噪比(SNR)和对比噪声比(CNR),并对三种图像质量进行主、客观评分。结果磨玻璃样病变与实性病变在混合能量、最佳单能量和水基图上SD值、SNR和CNR差异有统计学意义(各P<0.05)。磨玻璃样病变在水基图上主观评分最高,差异有统计学意义(P<0.05);实性病变在最佳单能量图和水基图主观评分高于混合能量图(各P<0.05)。结论能谱CT多参数成像通过重建最佳单能量图像和水基物质图像,可有效降低图像背景噪声,提高CT图像质量,尤其水基物质图像对提高COVID-19磨玻璃样病变图像质量的优势更明显,可在临床推广应用。Objective To explore the value of multi-parameter imaging in spectral CT for improving the image quality of Corona Virus Disease 2019.Methods A total of 53 ground-glass opacity(GGO)lesions and 23 solid lesions on the non-enhanced chest spectral CT images of 15 confirmed COVID-19 patients were retrospectively analyzed.Three sets of spectral CT images were selected for analysis,including conventional 120kVp-like CT image,monochromatic CT image with the highest contrast-noise ratio(CNR)(Opt-keV image)and the water-based material decomposition(MD)images.CT value(water density)of lesions and standard deviation(SD)of normal lung tissue were measured to calculate signal noise ratio(SNR)and contrast noise ratio(CNR)for lesions,and subjective image quality evaluation was performed.Results The water-based MD images provided the highest subjective score for GGO than the other CT images(all P<0.05).For solid lesions,the water-based MD images and Opt-keV images had similar subjective scores(all P>0.05)and were both higher than the 120kVp-like images(all P<0.05).The differences in SD values,SNR and CNR between GGO lesions and solid lesions on the 120kVp-like images,Opt-keV images and the water-based MD images were statistically significant(all P<0.05).Conclusion The spectral CT multi-parameter analysis can effectively reduce image noise through selecting Opt-keV CT image and using water-based MD image to improve lesion CNR.Water-based MD image has more obvious advantages in improving the image quality of COVID-19 ground-glass lesions.

关 键 词：能谱 计算机体层摄影术 新型冠状病毒肺炎 图像质量 

分 类 号：R511[医药卫生—内科学] R814.42[医药卫生—临床医学]