摘要
目的 探讨80 kVp条件下深度学习重建(DLR)在头颈心血管联合扫描中的应用价值。方法 前瞻性纳入2024年7月—2025年4月拟行头颈血管及冠状动脉成像的40例患者,采用80 kVp管电压结合可变螺距模式完成头颈心血管联合检查。采用DLR和混合迭代重建(HIR)算法重建头颈以及冠状动脉的图像。分别测量并计算颈总动脉(CCA)、颈内动脉(ICA)、大脑中动脉(MCA)、右冠状动脉(RCA)、左前降支(LAD)及左回旋支(LCX)的图像CT值、噪声值、信噪比(SNR)和对比噪声比(CNR)等客观指标。由两名经验丰富的放射科医师采用5分制Likert量表对图像进行主观评分。结果 DLR组与HIR组LCX的CT值[(593.59±152.38)HU比(590.03±152.50)HU]无统计学差异(P>0.05)。然而,DLR组在冠状动脉(RCA、LAD、LCX)与头颈部血管(CCA、ICA、MCA)的其他关键指标(包括CT值、标准差、SNR及CNR)均显著优于HIR组,差异均具有统计学意义(P<0.05)。此外,两名医师对DLR组的主观评分均为(4.65±0.58)分,显著高于对HIR组的评分[分别为(4.18±0.64)分、(4.25±0.63)分],差异均具有统计学意义(P<0.05),医师评分一致性高度一致(Kappa=0.757)。DLR组的噪声、SNR、CNR及主观图像质量评分均优于HIR组,差异均具有统计学意义(P<0.05)。结论 在80 kVp条件下的头颈心血管联合扫描中,DLR较HIR能更有效地提升图像质量。
Objective To explore the application value of deep learning reconstruction(DLR)in combined head-neck and cardiovascular computed tomography angiography(CTA)at 80 kVp.Methods A total of 40 patients scheduled for combined head-neck and cardiovascular CTA from July 2024 to April 2025 were prospectively enrolled.The CTA was performed using an 80 kVp tube voltage combined with a variable helical pitch mode.Images of the head-neck vessels and coronary arteries were reconstructed using DLR and hybrid iterative reconstruction(HIR).Objective metrics including CT values,noise,signal-to-noise ratio(SNR),and contrast-to-noise ratio(CNR)were determined for the common carotid artery(CCA),internal carotid artery(ICA),middle cerebral artery(MCA),right coronary artery(RCA),left anterior descending artery(LAD),and left circumflex artery(LCX).Two experienced radiologists evaluated the image quality subjectively using a 5-point Likert scale.Results There was no significant difference(P>0.05)in the CT values of the LCX between the DLR[(593.59±152.38 HU)]and HIR[(590.03±152.50 HU)]groups.The CT value,standard deviation,SNR and CNR of coronary arteries(RCA,LAD,LCX)and head-neck vessels(CCA,ICA,MCA)in the DLR group were significantly superior to those in the HIR group(P<0.05).The subjective scores of the DLR group(4.65±0.58 for both radiologists)were significantly higher(both P<0.05)than those of the HIR group(4.18±0.64,4.25±0.63)with good scoring consistency(Kappa=0.757).Conclusion DLR improves image quality compared to HIR in combined head-neck and cardiovascular CTA at 80 kVp.
作者
何小龙
盛杰鑫
刘亚良
郑超
陈堃
邱强
徐敏
张海为
何秋红
蔡勇智
杨琳
HE Xiaolong;SHENG Jiexin;LIU Yaliang;ZHENG Chao;CHEN Kun;QIU Qiang;XU Min;ZHANG Haiwei;HE Qiuhong;CAI Yongzhi;YANG Lin(Department of Radiology,Hanzhong Central Hospital,Shaanxi 723011,China)
出处
《影像诊断与介入放射学》
2026年第1期28-34,共7页
Diagnostic Imaging & Interventional Radiology
基金
2022年陕西省卫生健康科研基金资助项目(2022B011)。
关键词
深度学习重建
CT血管成像
冠状动脉血管造影
图像质量
80
kVp
Deep learning reconstruction
Computed tomography angiography
Coronary computed tomography angiography
Image quality
80 kVp
