摘要
目的探讨深度学习图像重建(DLIR)算法联合低管电压、低对比剂剂量(双低)在肺动脉CT血管造影中的可行性。方法60例临床拟诊为肺动脉栓塞(PE)的患者随机分成两组,每组30例,行CT肺动脉血管造影(CTPA)检查。常规组为A组,扫描管电压为100 kVp,对比剂用量为45 ml,采用ASIR-V50%重建;双低组扫描管电压为70 kVp,对比剂用量为35 ml,采用不同水平的DLIR重建,包括B组(DLIR-L)、C组(DLIR-M)、D组(DLIR-H)。比较双低组和常规组辐射剂量及图像质量。结果双低组和常规组的有效辐射剂量(ED)分别为(2.05±0.76)mSv、(2.56±0.68)mSv,差异有统计学意义(P<0.05)。双低组(B、C、D组)肺动脉主干、右肺上叶主干、右肺中叶内侧段、右肺下叶内侧基底段肺动脉CT值及右肺中叶内侧段、右肺下叶内侧基底段肺动脉和肺静脉CT值差值较常规组(A组)明显升高,差异有统计学意义(P<0.05)。双低组(B、C、D组)噪声较常规组(A组)降低,C组、D组较常规组(A组)差异有统计学意义(P<0.05),双低组内随着DLIR重建水平的增加,图像噪声递减。双低组(B、C、D组)SNR、CNR较常规组升高,C组、D组较常规组(A组)差异有统计学意义(P<0.05),双低组内随着DLIR重建水平的增加,SNR、CNR升高。双低组内随着DLIR重建水平的增加,主观评分增高,D组4分所占比例最多;A组和C组、A组和D组、B组和D组差异均有统计学意义(P<0.05)。结论DLIR算法联合低电压、低对比剂剂量行CTPA获得图像质量优于常规组,能满足临床诊断需求,同时降低了辐射剂量及对比剂用量。
Objective To explore the feasibility of deep learning image reconstruction algorithm combined with low tube voltage and low volume contrast dose in computed tomography pulmonary angiography.Methods Sixty patients with suspected pulmonary embolism were randomly divided into two groups,30 in each group,and received computed tomography pulmonary angiography.The conventional group was group A,the scanning tube voltage was 100kVp,the dosage of contrast agent was 45ml,and the ASIR-V50%reconstruction was adopted.The voltage of scanning tube in double-low group was 70kVp,and the dosage of contrast agent was 35ml,DLIR reconstruction was performed at different levels,including group B(DLIR-L),Group C(DLIR-M),and Group D(DLIR-H).Compare the radiation dose and image quality of double low group and conventional group.Results The effective radiation dose of double-low group and conventional group were(2.05±0.76)mSv,(2.56±0.68)mSv,respectively,and the difference was statistically significant(P<0.05).Compared with conventional group(A),CT values of pulmonary artery trunk,right upper lobe trunk,right middle lobe medial segment,right lower lobe medial basal segment,and CT value difference between arterial and venous of right middle lobe medial segment and right lower lobe medial basal segment in double-low group(group B,C and D)were significantly higher,with statistical significance(P<0.05).The noise of double-low group(group B,C and D)was lower than that of the conventional group(A),and the noise of group C and D was statistically significant compared with that of the conventional group(P<0.05),with the increase of DLIR reconstruction level in double-low group,the image noise decreased.SNR and CNR in double-low group(group B,C and D)were higher than that in conventional group(group A),the SNR、CNR of group C and D was statistically significant compared with that of the conventional group(P<0.05),with the increase of DLIR reconstruction level in double-low group,SNR and CNR were increased.With the increase of DLIR reconstruction level in double-low group,subjective score increased,and group D accounted for the largest proportion of 4 points,group A and C,group A and D,group B and D were statistically significant(P<0.05).Conclusion DLIR algorithm combined with low tube voltage and low volume contrast dose CTPA to obtain better image quality than the conventional group,which can meet the needs of clinical diagnosis and reduce the radiation dose and contrast agent dosage.
作者
杨彦兵
田艳鑫
王泽润
朱丽娟
阮小伟
于梓婷
汪芳
杨利莉
YANG Yanbing;TIAN Yanxin;WANG Zerun(People's Hospital of Ningxia Hui Autonomous Region(First Affiliated Hospital of Northwest National University),Yinchuan 750001,P.R.China)
出处
《临床放射学杂志》
北大核心
2023年第3期493-497,共5页
Journal of Clinical Radiology
基金
宁夏回族自治区重点研发(一般)项目(编号:2019BEG03046、2021BEG03092)
西北民族大学2019年中央高校基本科研项目(编号:31920190184)。
关键词
肺动脉CT血管造影
深度学习图像重建
低管电压
低对比剂
辐射剂量
Computed tomography pulmonary angiography
Deep learning image reconstruction
Low tube voltage
Low contrast agent
Radiation dose
