探讨不同管电压下DLIR重建算法对冠状动脉CTA图像质量的影响  

Effects of DLIR reconstruction algorithm on coronary CTA image quality under different tube voltages

作　　者：朱丽娟 马瑞 沈云 汪芳[1] 杨彦兵 曹永佩 杨利莉[1] 吴小红 ZHU Lijuan;MA Rui;SHEN Yun;WANG Fang;YANG Yanbing;CAO Yongpei;YANG Lili;WU Xiaohong(People’s Hospital of Ningxia Hui Autonomous Region,Yinchuan 750002,China;GE Computer Tomography Imaging Research Center of China,Shanghai 200120,China)

机构地区：[1]宁夏回族自治区人民医院,宁夏银川750002 [2]GE(中国)CT影像研究中心,上海200120

出　　处：《宁夏医学杂志》2023年第9期787-791,F0003,共6页Ningxia Medical Journal

基　　金：宁夏重点研发一般项目(2019BEG03046);宁夏重点研发项目(2021BEG03092);宁夏自然科学基金(2020AAC03346)。

摘　　要：目的DLIR深度学习图像重建技术与常规重建法(DLIR)对CCTA图像质量的影响。方法选取行冠状动脉CT血管成像的患者60例,根据扫描的管电压的不同分为A(70 kVp)、B(120 kVp)2组,每组患者各30例,均使用自适应迭代重建算法(ASiR-V40%)以及低级(DLIR-L)、中级别(DLIR-M)和高级别(DLIR-H)的深度学习算法重建原始数据。对比评价2组患者的主客观图像质量和辐射剂量。结果客观图像质量评分中A、B 2组间SNR的差异无统计学意义(P>0.05);在ASiR-V40%重建图像中,2组间CNR的差异无统计学意义(P>0.05),而在DLIR重建图像中,A组CNR较B组高(P<0.05)。主观图像质量评分中医师甲、医师乙两名观察者评分结果显示随着深度学习图像重建技术(DLIR)降噪级别的增高而升高(P<0.05),ASiR-V40%与DLIR-L主观图像评分没有明显差异;2组图像主观评分间差异无统计学意义(P>0.05)。A组的有效辐射剂量较B组降低了13.38%,且差异具有统计学意义(P<0.05)。对比剂的用量A组较B组降低25.0%。结论70 kVp管电压联合深度学习图像重建技术(DLIR)能提升CCTA图像的整体质量,降低辐射剂量及对比剂用量,具有较高的临床应用价值。Objective To explore the impact of DLIR deep learning image reconstruction technology and conventional reconstruction method(DLIR)on CCTA image quality.Methods Sixty patients who underwent Coronary CT Angiography(CCTA)in our hospital were selected.According to the different scanned tube voltages,and the patients were divided into two groups,A(70kVp)and B(120 kVp),with 30 patients in each group.The adaptive iterative reconstruction algorithm(ASiR-V40%),low-grade(DLIR-L),medium-grade level(DLIR-M)and high level(DLIR-H)deep learning algorithms was used to reconstruct raw data.Compared and evaluated the subjective and objective image quality and radiation dose of two groups of patients.Results In objective image quality score,there was no significant difference in SNR between groups A and B(P>0.05).In the ASiR-V40%reconstructed images,there was no significant difference in CNR between the two groups(P>0.05),while in the DLIR reconstructed images,the CNR of group A was higher than that of group B(P<0.05).In subjective image quality score:the scores of two observers,physician A and physician B,showed an increase with the increase of deep learning image reconstruction technology(DLIR)noise reduction level(P<0.05),and there was no significant difference between ASiR-V40%and DLIR-L subjective image score.There was no statistically significant difference in subjective scores between the two groups of images(P>0.05).Radiation dose comparison:compared to group B,the effective radiation dose of group A decreased by 13.38%,and the difference was statistically significant(P<0.05).Comparison of the dosage of contrast agents:compared to group B,the dosage of contrast agents of group A decreased by 25.0%.Conclusion The combination of 70 kVp tube voltage and deep learning image reconstruction technology(DLIR)can improve the overall quality of CCTA images,reduce radiation dose and contrast agent dosage,and have high clinical application value.

关 键 词：冠状动脉CT血管成像 图像质量 管电压 辐射剂量 深度学习图像重建 

分 类 号：R816[医药卫生—放射医学]