放射学实践2024,Vol.39Issue(10) :1361-1366.DOI:10.13609/j.cnki.1000-0313.2024.10.013

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

曹建新 朱卓 张清 韩锦涛 陈颖 许炜
放射学实践2024,Vol.39Issue(10) :1361-1366.DOI:10.13609/j.cnki.1000-0313.2024.10.013

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

曹建新 1朱卓 1张清 1韩锦涛 2陈颖 1许炜1
扫码查看

作者信息

  • 1. 430081 武汉,武汉科技大学附属普仁医院放射科
  • 2. 201800 上海,上海联影医疗科技股份有限公司
  • 折叠

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.

Abstract

Objective:To explore the application value of dramatic reducing radiation dose of CT pulmonary angiography(CTPA)using artificial intelligence model-based iterative reconstruction(AI-IR)combined with 70kV ultra-low tube voltage setting.Methods:A total of 100 patients with suspec-ted pulmonary embolism or follow-up of pulmonary artery embolism were underwent CTPA and ran-domly 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 pa-tients,10 patients were excluded because of respiratory motion artifacts or massive pleural effusion.Fi-nally,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 in group A and B(P<0.001).The subjective scores were 4.5±0.6,4.1±0.5 and 4.9±0.3 for the HIR images in group A,the HIR images and AIIR images in group B,respectively;the subjec-tive scores for the AIIR images in group B were significantly higher than those for the HIR images in group A and B(P<0.001).Conclusion:The radiation dose of CTPA can be reduced significantly using AIIR combined with 70kV ultra-low tube voltage setting while image quality is better than the HIR image using 100kV tube voltage.

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

Key words

CT angiography/Artificial intelligence/Image processing,computer-assisted/Im-age enhancement/Radiation dosage/Tomography,X-ray computed

引用本文复制引用

基金项目

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)

出版年

2024
放射学实践
华中科技大学同济医学院

放射学实践

CSTPCDCSCD北大核心
影响因子:1.08
ISSN:1000-0313
参考文献量19
段落导航相关论文