联合人工智能模型迭代重建和70kV超低管电压大幅度降低肺动脉CTA辐射剂量的研究  

Study of dramatic reducing radiation dose of CT pulmonary angiography using artificial intelligence model-based iterative reconstruction with 70kV ultra-low tube voltage setting

作　　者：曹建新 朱卓 张清[1] 韩锦涛 陈颖[1] 许炜[1] CAO Jian-xin;ZHU Zhuo;ZHANG Qing(Department of Radiology,Puren Hospital Affiliated to Wuhan University of Science and Technology,Wuhan 430081,China)

机构地区：[1]武汉科技大学附属普仁医院放射科,武汉430081 [2]上海联影医疗科技股份有限公司,上海201800

出　　处：《放射学实践》2024年第10期1361-1366,共6页Radiologic Practice

基　　金：武汉科技大学职业危害识别与控制湖北省重点实验室开放基金项目(OHIC2021G04)。

摘　　要：目的:探讨联合人工智能模型迭代重建(AIIR)和70kV超低管电压在大幅度降低肺动脉CTA(CTPA)辐射剂量中的应用价值。方法:100例疑似肺动脉栓塞或肺动脉栓塞复查的患者随机分成A组(100 kV管电压)和B组(70 kV管电压)行CTPA检查,A组采用混合迭代重建(HIR),B组采用HIR和AIIR。评估A、B两组辐射剂量。评估A组HIR图像、B组HIR图像和AIIR图像背景噪声、肺动脉信噪比(SNR)和对比噪声比(CNR),并采用5分法对CTPA图像质量进行评分。结果:100例患者中10例存在呼吸运动伪影或大量胸腔积液被排除,最终A组和B组分别有50例和40例患者纳入。A、B两组有效辐射剂量(ED)分别为(4.3±0.8)mSv和(1.3±0.2)mSv,与A组相比B组ED下降69.8%(P<0.001)。A组HIR图像、B组HIR和AIIR图像背景噪声分别为(19.2±3.7)HU、(28.1±4.4)HU和(13.7±6.5)HU;B组AIIR图像背景噪声显著低于A组HIR图像和B组HIR图像(P<0.001)。A组HIR图像上右肺动脉信噪比(SNR)和对比噪声比(CNR)分别为21.5±5.9和18.8±5.6,左肺动脉SNR和CNR分别为21.9±6.2和19.2±5.9;B组HIR图像上右肺动脉SNR和CNR分别为19.4±3.7和17.2±3.5,左肺动脉相应数值分别为19.4±3.5和17.2±3.3;B组AIIR图像上右肺动脉SNR和CNR分别为45.7±14.5和40.9±13.4,左肺动脉相应数值分别为46.0±14.6和41.1±13.6;B组AIIR图像上右肺动脉及左肺动脉SNR和CNR均明显高于A组HIR图像和B组HIR图像(P<0.001)。A组HIR图像、B组HIR图像和B组AIIR图像主观评分分别为4.5±0.6、4.1±0.5和4.9±0.3,B组AIIR图像主观评分显著高于A组和B组HIR图像(P<0.001)。结论:联合AIIR和70kV超低管电压在大幅度降低CTPA辐射剂量的同时,图像质量相对于100kV管电压HIR图像更优。Objective:To explore the application value of dramatic reducing radiation dose of CT pulmonary angiography(CTPA)using artificial intelligence model-based iterative reconstruction(AIIR)combined with 70kV ultra-low tube voltage setting.Methods:A total of 100 patients with suspected pulmonary embolism or follow-up of pulmonary artery embolism were underwent CTPA and randomly divided into group A(100kVp)and group B(70kVp),hybrid iterative reconstruction(HIR)was used in group A,and HIR and AIIR were used in group B.Radiation doses were evaluated for group A and B.Background noise,signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR)of pulmonary artery on the HIR images in group A and those on the HIR and AIIR images in group B were evaluated,image quality of CTPA was graded on a 5-point scale.Results:Among the 100 patients,10 patients were excluded because of respiratory motion artifacts or massive pleural effusion.Finally,50 and 40 patients were included in group A and group B,respectively.The effective doses(ED)were(4.3±0.8)mSv in group A and(1.3±0.2)mSv in group B,compared to group A,the ED was decreased by 69.8%in group B(P<0.001).The background noise was(19.2±3.7)HU,(28.1±4.4)HU and(13.7±6.5)HU on the HIR images in group A,the HIR and AIIR images in group B,respectively;the background noise on the AIIR images in group B was significantly lower than that on the HIR images in group A and B(P<0.001).On the HIR images in group A,the SNR and CNR of right pulmonary were 21.5±5.9 and 18.8±5.6,and those of left pulmonary were 21.9±6.2 and 19.2±5.9,respectively;on the HIR images in group B,the SNR and CNR of right pulmonary were 19.4±3.7 and 17.2±3.5,and those of left pulmonary were 19.4±3.5 and 17.2±3.3,respectively;on the AIIR images in group B,the SNR and CNR of right pulmonary were 45.7±14.5 and 40.9±13.4,and those of left pulmonary were 46.0±14.6 and 41.1±13.6,respectively;the SNR and CNR of right and left pulmonary on AIIR images in group B were all significantly higher than those on HIR images

关 键 词：CT血管造影术 人工智能 图像处理 计算机辅助 图像增强 辐射剂量 体层摄影术 X线计算机 

分 类 号：R-05[医药卫生] R814.3R814.4R814R144