Objective:This study was aimed to investigate the impact of deep learning image reconstruction（DLIR） algorithm on quantitative analysis and image quality of pulmonary CT by comparing with filtered back projection（FBP） and adaptive statistical iterative reconstruction veo（ASIR-V）.Methods:Non-contrastimages of 46 subjects who underwent lung CT screeningwere collected retrospectively and reconstructed using FBP,ASIR-V with 50% and 100% hybrid iterative reconstruction blending weights（BW）,and different levels（low, medium and high） of deep learning-based reconstruction algorithm（named DLIR-L,DLIR-M and DLIR-H,respectively）.Quantitative parameters were compared including average CT values and volumes of solid（VS） and non-solid areas（VNS） of each nodule, lung volume（LV）,emphysema index（EI）,and luminal area（LA）,airway diameter（AD）,wall area（WA） and wall thickness（WT） of the right upper lobular apical segment, CT value and its standard deviation（SD,represented as image noise） of the descending aorta and CT value and its SD of the air inside the trachea at the level 1cm above the tracheal bifurcation.Image quality and noise in different reconstruction algorithms were evaluated by two radiologists using a 5-point scale（1:very poor; 2:poor; 3:qualified; 4:good; 5:excellent） and compared statistically.Results:Compared with FBP,the image noise was significantly reducedby using different level DLIR algorithms and ASIR-V algorithms with 50% and 100% BW（all P＜0.001）.Compared with FBP,the image noise was significantly reduced by using 50%-ASIR-V,DLIR-L and DLIR-M algorithms, DLIR-H algorithms（P＜0.001）.Different algorithms had a significant effect on emphysema index（P＜0.001）.There were no significant differences in total lung volume, airway area, airway wall area, average airway wall thickness, airway diameter, CT values of air and descending aorta, and nodule volume（VS and VNS） among different reconstructions（all P＞0.05）.DLIR-H algorithm had higher subjective evaluating score of image quality than that of the other algorithms（all P＜0.05）.And the subjective score of image noise of DLIR-H algorithm was higher than that of FBP,50%-ASIR-V,DLIR-L and DLIR-H algorithms（all P＜0.05）.Conclusion:Compared with FBP,DLIR and ASIR-V algorithm could reduce the image noise and improve the image quality without affecting the quantitative results of the pulmonary structures and lung nodules.Compared with ASIR-V,DLIR-H algorithm can significantly reduce the image noise and improve the image quality for clinical diagnosis.

目的 探讨胸腺原发黏膜相关淋巴组织淋巴瘤(MALToma)的CT表现,以期提高对该病的认识.方法回顾性分析8例经病理证实的胸腺MALToma患者的临床及CT特征并复习相关文献.结果 8例患者中,5例为囊实性铸型生长肿块,内可见多发边界清晰的类圆形囊变区域,定义为"瑞士奶酪征",3例为实性类圆形小结节;增强后实性成分均中度至明显强化,平均强化幅度为35 HU.8例均无纵隔肿大淋巴结,无心脏大血管包绕或侵犯.7例合并干燥综合征.结论 胸腺MALToma的CT表现有一定的特征性,且多数伴有干燥综合征,结合患者临床病史有可能做出准确的术前诊断.

目的 评价自适应统计迭代重组算法(ASIR-V)混合权重对冠状动脉CT血管造影(CCTA)量化分析结果的影响.方法48例CCTA扫描原始数据分别利用滤波反投影法(FBP)、ASIR-V20％、ASIR-V40％、ASIR-V60％、ASIR-V80％混合迭代算法重组.测量升主动脉CT值平均值及标准差、狭窄处管腔最小径、狭窄程度、管腔体积、管腔面积、斑块体积、斑块面积及冠状动脉周围脂肪(PCAT)CT值等.比较不同迭代混合权重对上述测量结果的影响.结果 随着 ASIR-V 比例的提高(FBP,ASIR-V20％,ASIR-V40％,ASIR-V60％,ASIR-V80％),图像噪声逐渐降低[升主动脉 CT值标准差:(39.9±6.8)H1U,(35.5±6.2)H1U,(31.2±5.7)HU,(27.0±5.3)H1U,(23.0±5.0)HU](P＜0.05),其他量化结果的差异无统计学意义.结论ASIR-V混合权重影响CCTA图像噪声水平,对量化分析结果的影响无统计学意义.

目的 评价利用256排探测器、16 cm Z轴覆盖范围CT机行自由呼吸冠状动脉血管成像(CCTA)的可行性.方法 将616例CCTA检查者随机分为两组,A组(325例)屏气扫描,B组(291例)自由呼吸扫描.比较两组图像质量评分、信噪比、辐射剂量.A组22例,B组24例,在CCTA检查后两周内接受经导管冠状动脉造影(ICA).以ICA结果作为金标准,比较两种扫描方法诊断血管狭窄(＞50％)的敏感度、特异度.结果 两组心率[A组(70.8±13.8)bpm;B组(70.7±13.2) bpm]、图像质量评分(A组1.49±0.62;B组1.53±0.67)、有效辐射剂量、信噪比等差异均无统计学意义(P＞0.05).基于节段分析,A组敏感度82.1％、特异度96.8％;B组敏感度82.2％、特异度96.6％,差异均无统计学意义(P＞0.05).结论 利用16 cm Z轴覆盖范围CT机行自由呼吸CCTA,图像质量、诊断准确率与屏气扫描无差异.

目的 评价利用256排探测器CT行自由呼吸冠状动脉钙化积分扫描的可行性.方法 100例钙化积分扫描受检者随机分为两组:A组(50例)屏气扫描,B组(50例)自由呼吸扫描.比较两组受检者心率、主动脉平均CT值、CT值标准差、有效剂量、斑块运动伪影主观评分.比较两组钙化积分,包括Agatston积分、体积积分(mm3)、质量积分(mg)及基于Agatston积分的心血管事件危险分层.结果 两组受检者一般资料、主动脉平均CT值、CT值标准差、有效剂量的差异无统计学意义(P＞0.05).两组间Agatston积分(A组188.10±445.84;B组152.26±321.86)、质量积分(A组24.64±54.40 mg;B组22.60±51.19 ag)、体积积分(A组73.72±159.49mm3;B组63.58±120.40mm3)及危险分层差异无统计学意义(P＞0.05).结论 利用256排探测器CT行自由呼吸冠状动脉钙化积分扫描,积分计算结果、辐射剂量等与屏气扫描无差别.

目的:评价混合迭代重建算法(ASIR-V)混合权重对心脏CT冠状动脉钙化积分及危险分层计算结果的影响.方法:100例心电门控心脏CT钙化积分扫描原始数据利用不同混合权重ASIR-V算法重建,包括ASIR-V 0％即(FBP)、ASIR-V 20％、ASIR-V 40％、ASIR-V 60％、ASIR-V 80％、ASIR-V 100％.比较不同算法所得图像升主动脉CT值平均值及标准差、钙化斑块最大CT值、钙化积分(Agatston积分、质量积分、体积积分)及基于Agatston积分危险分层.结果:随ASIR-V混合权重提高升主动脉CT值平均值及冠脉钙化斑块最大CT值变化差异无统计学意义(P>0.05),升主动脉CT值标准差、3种钙化积分逐渐降低且差异有统计学意义(P<0.05).Agatston积分分别为170.2±387.3、168.6±385.3、166.4±381.9、164.6±379.4、162.9±376.7、161.1±374.1(P=0.000).质量积分分别为23.62±52.56、23.49±52.23、23.32±52.30、23.20±53.10、23.15±52.83、22.99±52.41mg(P=0.000).体积积分分别为68.7±140.7、67.8±139.8、67.0±139.6、65.7±137.2、64.8±135.9、64.0±135.0mm3(P=0.000).ASIR-V导致10位受检者危险分层发生改变,其中8人从低危险降至极低危险,2人从中等危险降至低危险.结论:ASIR-V重建算法混合权重的提高可降低图像噪声水平,降低钙化积分,可能造成危险分层降低.

目的 评价迭代重建算法(ASIR-V)对心脏CT冠状动脉Agatston钙化积分及风险分级的影响.方法回顾性分析100例冠状动脉钙化积分CT扫描.原始数据利用不同混合百分比的ASIR-V算法重建,包括FBP即ASIR-V 0％、ASIR-V 20％、ASIR-V 40％、ASIR-V 60％、ASIR-V 80％、ASIR-V 100％.比较各组图像升主动脉 CT 值的最大值、平均值及标准差、钙化斑块最大CT值、Agatston积分及基于Agatston积分的风险分级.结果 随着ASIR-V百分比的增加,升主动脉CT值的平均值及钙化斑块最大CT值变化差异无统计学意义(P＞0.05),主动脉CT值最大值、标准差、钙化积分逐渐降低且差异有统计学意义(P＜0.05).Agatston积分分别为170.2±387.3、168.6±385.3、166.4±381.9、164.6±379.4、162.9±376.7、161.1±374.1.升主动脉 CT 值标准差分别为24.6±2.5、21.8±2.3、19.0±2.1、16.4±2.0、13.8±1.7、11.6±1.6 HU.10例受检者的风险分级发生改变,其中8例从低级降至极低级,2例从中级降至低级.结论 ASIR-V可降低图像噪声水平,降低钙化积分并可能造成风险分级的降低.计算钙化积分及风险分级时须注意图像重建算法.

目的:确定冠状动脉CT血管成像(CCTA)最佳重建期相,优化采集时间窗.评价优化采集时间窗对图像质量及辐射剂量的影响.方法:400例受检者(A组)行全心动周期(0～100％ RR间期)采集时间窗冠状动脉CT血管成像,确定最佳重建期相.优化采集时间窗设置为最佳重建期相的95％置信区间,即平均值±标准差×2.比较优化采集时间窗扫描模式(B组,400例)与全期相扫描模式(A组)的心率、信噪比、图像质量主观评分及辐射剂量,并进行统计学分析.结果:心率＜61 bpm者纳入低心率组(A1、B1);心率介于61～75 bpm者纳入中心率组(A2、B2);心率＞75 bpm者纳入高心率组(A3、B3).A1组最佳重建期相(76±3)％RR间期(95％置信区间70％～82％).A2组最佳重建期相(44±3)％及(76±3)％ RR间期(95％置信区间:38％～50％及70％～82％).A3组最佳重建期相(48±5)％RR间期(95％置信区间38％～58％).A、B两组的心率[A组为(70.54±13.06) bpm,B组为(70.38±13.05) bpm)]、图像质量主观评分[A组为(1.52±0.65)分,B组为(1.51±0.63)分]、信噪比(A组为15.76±3.70,B组15.66±4.59)差异均无统计学意义(P值均＞0.05).A1组[(36.73±14.39) mGy]与B1组[(7.13±2.53) mGy]、A2组[(27.57±11.69) mGy]与B2组[(12.13±3.39) mGy]、A3组[(25.55±10.33) mGy]与B3组[(7.74±3.13) mGy]之间的容积CT剂量指数(CTDIvol]差异均有统计学意义(P值均＜0.05).对于低、中、高心率受检者,优化采集时间窗扫描模式与全心动周期扫描模式相比,辐射剂量分别降低了81％0、56％、70％.结论:慢心率者冠脉最佳成像时间位于舒张中末期,高心率者位于收缩末期;中心率者位于收缩末期或舒张中末期.优化采集时间窗,可以在保证图像质量的前提下缩短曝光时间从而降低辐射剂量.

Objectives: To evaluate the feasibility of coronary computed tomography angiography (CCTA) in patients with free-breathing using 16-cm z-coverage CT with motion correction algorithm. Methods: 616 patients underwent CCTA without heart rate control. 325 examinations were performed during breath-holding (group A), and the remaining 291 were performed during free-breathing (group B). The image quality scores were defined as 1 (excellent), 2 (good), 3 (adequate), and 4 (poor). 22 patients in group A and 24 in group B underwent invasive coronary angiography (ICA) after CCTA within two weeks. The image quality score, diagnostic accuracy using ICA as reference, signal-to-noise ratio (SNR), and effective dose (ED) were compared between the two groups. Results: Mean heart rate during scanning was 70.8 +/- 13.8 bpm in group A and 70.7 +/- 13.2 bpm in group B (P = .950). No significant differences were observed in SNR and image quality score (1.49 +/- 0.62 vs. 1.53 +/- 0.67; P = .647) between the breath-holding and free-breathing groups. ED (1.99 +/- 0.83 mSv vs. 2.01 +/- 0.88 mSv) was not significantly different between the two groups (P = .975). In a segment-based analysis, the sensitivity, specificity and diagnostic accuracy in the detection of coronary stenosis of more than 50% were 82.1%, 96.8% and 92.2%, respectively in the breath-holding group and 82.2%, 96.6% and 92.2%, respectively in the free-breathing group with no significant differences for these parameters between the two groups. Conclusions: CCTA for patients without heart rate control and during free-breathing using 16-cm z-coverage CT with motion correction algorithm showed no significant difference in image quality and diagnostic performance compared with CCTA during breath-holding.

目的 探讨冠状动脉CT成像(CCTA)最佳重建期相随心率变化的规律.依据结果设置最佳采集时间窗,并计算其降低辐射剂量的能力.方法 回顾性分析2016年11月至2017年10月以心功能分析为检查目的接受全心动周期全剂量CCTA的616例患者资料.由2名诊断医师确定最佳重建期相,分析最佳重建期相随心率变化的规律.依据此结果设定根据心率决定的窄采集时间窗模式.采集时间窗位于最佳重建期相的平均值±标准差×2,即95％置信区间.计算此采集模式较全心动周期采集模式降低辐射剂量的程度.结果 616例受检者,心率(71±14)次/分(42～149次/分).低心率组(＜61次/分)最佳重建期相:(76±3)％RR间期,95％置信区间:70％～ 82％ RR间期;中心率组(61～ 75次/分)最佳重建期相以60％ RR间期为界位于(44±3)％及(76±3)％ RR间期,95％置信区间:38％～50％ RR间期及70％～82％ RR间期;高心率组(＞75次/分)最佳重建期相:(48±5)％RR间期,95％置信区间:38％～58％RR间期.以95％置信区间作为采集时间窗,较全心动周期采集模式降低CT-DIvol约80％(低心率)、56％(中心率)、70％(高心率).结论 慢心率者冠状动脉运动相对静止期位于舒张中末期;高心率者位于收缩末期;中心率者位于收缩末期或舒张中末期.以最佳重建期相95％置信区间作为采集时间窗,可以在保证图像质量的前提下缩短曝光时间从而降低辐射剂量.

目的:通过体模研究,评价双能量CT扫描获得的虚拟单色谱图像(VMS)联合金属伪影削减(MAR)算法抑制股骨柄假体金属伪影的效果.方法:将人工股骨柄假体用细绳悬挂在水箱中心.以不同管电压(80、100、120、140 kVp)单能量CT (SECT)及不同管电流(200、484 mA)下的双能量CT(DECT)扫描假体.扫描设备选择单球管,单层探测器CT扫描仪.利用单球管双电压快速切换的双能量扫描获得的数据重建40、90、140 keV单色谱图像(VMS),及经过金属伪影抑制算法处理的40、90、140 keV单色谱图像(VMS+MAR).对全部图像(16组)进行质量评价,包括伪影面积(AA),伪影指数(AI).结果:16组图像(140 kVp、120 kVp、100 kVp、80 kVp、40 keV 485 mA、40 keV 485 mA MAR、90 keV 485 mA、90 keV 485 mA MAR、140 keV 485 mA、140 keV 485 mA MAR、40 keV 200 mA、40 keV 200 mA MAR、90 keV 200 mA、90 keV 200 mA MAR、140 keV 200 mA、140 keV 200 mAMAR)伪影面积分别为787、1088、1499、2133、5664、650、665、172、279、157、7603、886、803、238、423、172 mm2,伪影指数分别为22.75、33.65、58.58、92.55、195.18、3.53、23.38、0.35、7.88、0.05、230.28、4.95、24.83、1.43、10.98、0.10 HU.对于单能量扫描,随着管电压的降低,伪影逐步加重.对于双能量扫描单色谱图像,管电流越低伪影越严重;光子能量越低伪影越严重.经过金属伪影削减算法处理的单能量图像,金属伪影明显减轻.结论:与单能量扫描相比,双能量CT扫描高光子能量单色谱图像联合MAR算法能有效抑制股骨柄假体金属伪影.光子能量越高,金属伪影越小.

Objective To investigate the impact of parameters setting on image quality and radiation dose of low-dose chest CT scanning when iCT tube current was set as 17 mAs.Methods Two hundred volunteers who accepted low-dose chest CT screening with tube current was set as 17 mAs were randomly divided into four groups according to the tube voltage and automatic exposure control (DoseRight) settings (group A:120 kV,DoseRight;group B:100 kV,DoseRight;group C:120 kV,fixed tube current;group D:100 kV,fixed tube current).Then hybrid iterative reconstruction (HIR:iDose4) and iterative model reconstruction (IMR) were performed on all scanned images of lung and mediastinal window algorithm,respectively.The effective dose and image quality of 4 groups were compared.Results There were significant differences of image quality score and effective dose among 4 groups.The image quality of group A and B were better than that of group C (P＜0.05),and of group D was the worst compared with the other three groups (P＜0.05).Image quality scores of the mediastinal window of group C and D and the lung window of group D at iDose4 reconstruction were lower.SNR value of group A,B,C,D decreased in order (P＜0.01).Group A had the highest effective dose (P＜0.01),while the effective dose of group B and group C had no significant difference (P=0.055),and group D had the lowest effective dose (P＜ 0.01).Conclusion When tube current was set as 17mAs at iCT,selecting 100 kV tube voltage,DoseRight and IMR reconstruction arithmetic can obtain satisfactory image quality and low radiation dose.

To evaluate the clinical value of virtual monochromatic spectral (VMS) images with metal artifact reduction (MAR) algorithm in dual-energy computed tomography (DECT)-guided microcoil localization of pulmonary nodules. Fifty-one patients underwent DECT after placement of microcoils on small pulmonary nodules before video-assisted thoracoscopic surgery (VATS). Optimal energy level (in keV) was defined as the level at which CT values of nodules were equivalent to those of 120 kVp images and with no serious metal artifacts. VMS images at optimal keV and at 50, 90,110, and 140keV with and without MAR were reconstructed. Image quality was scored using a 3-point scale: 1=excellent, minimal artifacts; 2=good, mild artifacts; and 3=poor, extensive artifacts. Image quality scores between the VMS-only and VMS + MAR groups were compared; 74 keV was found to be the optimal level for VMS images. The image quality of the VMS + MAR images at 74keV were significantly better than VMS-only images (1.35 +/- 0.59 vs 2.11 +/- 0.87, P=.005). There was no difference in image quality score among VMS + MAR images at 74 keV and higher energy levels. VMS images from DECT at 74 keV with MAR can reduce artifacts from microcoils and improve image quality for microcoil localization of pulmonary nodules.

Objective This study aimed to investigate the impact of hybrid iterative reconstruction blending percentage on quantitative analysis of pulmonary CT.Methods Pulmonary CT images were acquired from 52 subjects and reconstructed using different hybrid iterative reconstruction blending percentage including FBP,ASIR-V30％,ASIR-V50％,ASIRV70％ and ASIR-V100％.Quantitative analyses were compared among the five algorithms including emphysema index (EI),total lung volume,luminal area,wall area,wall thickness,diameter of airway and nodule volume.CT value,nodule volume and standard deviation (SD) were also measured and compared.Results EIs were found statistically different FBP (3.34 ％),ASIR-V30％ (2.13 ％),ASIR-V50％ (1.46％),ASIR-V70％ (0.98％)andASIR-V100％ (0.57 ％) (P =0.000).SDs decreased dramatically in the order of increasing ASIR-V percentage air in trachea:FBP (29.8 HU),ASIR-V30％ (25.5 HU),ASIR-V50％ (22.6 HU),ASIR-V70％ (19.6 HU),ASIR-V100％ (15.2 HU),descending aorta:FBP (35.2HU),ASIR-V30％ (27.6HU),ASIR-V50％ (22.7HU),ASIR-V70％ (18.0HU),ASIRV100％ (12.1HU) (P =0.000).There was no statistically significant difference in CT values,airway dimensions,nodule volumes and lung volumes among the different algorithms.Conclusion With increasing blending percentage of ASIR-V,EIs and SDs decreased significantly,while volumetric quantification of lung nodule and airway dimensions were not affected with statistically significant difference.

目的 探讨自由呼吸、不屏气状态下,利用256-MDCT行冠状动脉造影的可行性.方法 80例受检者随机分为两组:A组40例,扫描时须受检者屏气配合;B组40例,检查全过程无屏气口令,受检者自由呼吸.比较两组图像噪声水平、信噪比、主观评分(3分制:3分最佳,2分尚可,1分不能满足诊断)、辐射剂量.比较两组增强扫描前后心率变化程度.结果 两组间噪声水平、信噪比、辐射剂量均无统计学差异.增强扫描前与增强扫描中心率差异:A组(6.80±7.59)bpm,B组(3.03 ±2.58)bpm(P =0.012).图像质量主观评分:A组(2.55±0.64)分,B组(2.85 ±0.36)分(P=0.018).两种扫描模式相比,自由呼吸状态心率变化较小,检查效果更佳.结论 针对本研究中使用的设备及扫描方案,自由呼吸状态下行CCTA检查,效果优于屏气状态下检查.

Usually, coronary computed tomography angiography (CCTA) is performed during breath-holding to reduce artifact caused by respiration. The objective of this study was to evaluate the feasibility of free-breathing CCTA compared to breath-holding using CT scanner with wide detector. To evaluate the feasibility of CCTA during free-breathing using a 256-MDCT. In 80 patients who underwent CCTA, 40 were performed during breath-holding (group A), and the remaining 40 during free-breathing (group B). The quality scores for coronary arteries were analyzed and defined as: 3 (excellent), 2 (good), and 1 (poor). The image noise, signal-to-noise ratio and effective radiation dose as well as the heart rate variation were compared. The noise, signal-to-noise ratio, and effective radiation dose were not significantly different between the 2 groups. The mean heart rate variation between planning and scanning for group A was 7 +/- 7.6bpm, and larger than 3 +/- 2.6bpm for group B (P=0.012). Quality scores of the free-breathing group were better than those of the breath-holding group (group A: 2.55 +/- 0.64, group B: 2.85 +/- 0.36, P=0.018). Free-breathing CCTA is feasible on wide detector CT scanner to provide acceptable image quality with reduced heart rate variation and better images for certain patients.

Objective To discuss the application of adaptive statistic iterative reconstruction Veo(ASiR-V)in low-dose chest CT screening.Methods Totally 100 patients were categorized randomly into low-dose group(120kVp/30mAs)and ultra-low-dose group(120kVp/17.5 mA).Images were reconstructed with different ASiR-V blending percentages(0,30%,50%,70%,100%).Subjective image score,number of nodules and the standard deviation(SD)of CT value among different ASiR-V blending percentages,SD among different scores in ASiR-V50% were all compared.Results The scores of lung algorithm images with ASiR-V70% were higher than the others.The most nodules were detected with ASiRV70%.The scores for standard algorithm images with ASiR-V50% were higher than the others.SD values in subjective scoring 2,3,4and 5of ASiR-V50% were(24.20±1.01)HU,(20.99±2.32)HU,(19.54±2.16)HU and(16.90±2.60)HU respectively,which had statistically significant between any two(all P0.05).Conclusion 70% is the best blending percentage for lung algorithm and 50%for standard algorithm;20.99 HU of SD at ascending aorta is the criteria of images reconstructed by standard algorithm with ASiR-V50%for low-dose chest screening.