中文摘要：

目的 探讨能谱CT 单能量成像联合自适应统计迭代重建（ASiR）技术在兔正常肝脏增强扫描中的应用价值.方法 健康新西兰大白兔12 只,每只兔子使用CT 能谱成像（GSI）扫描模式行腹部双期增强扫描,在9 个单能量水平（40、50、60、65、70、75、80、90、110 keV）使用不同比例ASiR 混合算法（10%～100%,间隔10%）和滤波反投影（FBP）算法重建图像,得到A组和B 组图像.24 h 后对每只兔子行100 kV 管电压常规腹部双期增强扫描,亦使用不同比例ASiR混合算法（10%～100%,间隔10%）和FBP 算法重建图像,得到C 组和D 组图像.由两名放射学家对4 组门静脉期图像质量采用5 分评分制法进行主观评分,采用秩和检验比较,观察者间-致性采用Kappa 检验.测量4 组门静脉期图像的噪声、肝脏信噪比（SNR）和对比噪声比（CNR）,并测量最佳CNR 单能量水平,结果采用单因素方差分析及配对t 检验比较.结果 GSI 扫描组有效辐射剂量（2.76±0.23）mSv 低于100 kV 常规扫描组（3.21±0.28）mSv（t=2.787,P=0.018）.A 组图像中65～75 keV 单能量联合40% ASiR 时图像质量最好,主观评分最高;B 组图像中65～75 keV 单能量水平时图像质量最好;C 组图像中联合40% ASiR 时图像质量最好;4 组图像主观评分差异有统计学意义,A 组图像高于B、C、D 三组图像（P〈0.05）.总体上,随着单能量水平的升高及ASiR 比例的增加,图像噪声逐渐降低,在110 keV 联合100% ASiR 时图像噪声最低[（3.17±0.81）HU];70 keV 联合40% ASiR 时图像噪声[（6.41±0.67）HU]低于70 keV 联合FBP[（8.09±0.87）HU]、100 kV 联合40% AsiR[（9.56±0.79）HU]和100 kV 联合FBP[（12.07±1.07）HU]（均P〈0.05）.肝脏SNR 在9 个单能量水平中于70 keV 最高;联合ASiR 重建时,随着ASiR 比例的增加,SNR 逐渐升高,在70 keV 联合100% ASiR 时图像SNR 最高（14.48±0.56）;70 keV 联合40% ASiR 时肝脏SNR（13.78±0.58）高于70 keV 联合FBP（11.8

英文摘要：

ObjectiveTo investigate the impact of spectral CT monochromatic imaging and adaptive statistical iterative reconstruction techniques on the image quality in the portal venous phase rabbit normal liver.Methods12 New Zealand white rabbits underwent two times liver contrast enhanced CT with gemstone spectral imaging (GSI) and 100 kV tube voltage separately on the same GSI capable 64-slice CT scanner (Discovery CT 750 High Definition, GE Healthcare), and the time interval was 24 h. The first scan in 9 sets of monochromatic level (40, 50, 60, 65, 70, 75, 80, 90 and 110 keV) and the second scan in 100 kV tube voltage condition both underwent increasing contributions of ASiR (10%-90%, internal 10%) and FBP reconstruction. The objective image noises, signal-to-noise ratios (SNR) and contrast-to-noise ratios (CNR) were measured. Two radiologists independently and blindly assessed the subjective image quality. The results were analyzed using the pairedt-test, Wilcoxon signed rank sum test and mixed-effects model with Bonferronicorrection.Results The effective dose (ED) in the first GSI scan (2.76±0.28)mSv was lower than in the second 100 kV scan (3.21±0.32)mSv. At 65-75 keV combined with 40% ASiR yielded a maximal subjective image quality score. The objective image noise was the lowest at 110 keV with 100% ASiR (3.17±0.81 HU); and at 70 keV with 40% ASiR (6.41±0.67) HU was below at 70 keV with FBP (8.09±0.87) HU, 100 kV with 40% ASiR (9.56±0.79) HU and 100 kV with FBP (12.07±1.07) HU (allP<0.05). Liver SNR at 70 keV with 100% ASiR (14.48±0.56) reach the maximal; and at 70 keV with 40% ASiR (13.78±0.58) was above at 70 keV with FBP (11.82±0.73), 100 kV with 40% ASiR (9.23±0.54) and 100 kV with FBP (8.11±0.86)(allP<0.05). The optimal CNR was at 40-43 keV (58.33%) and 68-72 keV (41.67%); Liver CNR at 70 keV with 40% ASiR (4.84±0.17) was above at 70 keV with FBP (4.11±0.21), 100 kV with 40% ASiR (3.85±0.18) and 100 kV with FBP (3.17±0.25)(allP<0.05). Conclusion In abdominal enhanced CT of rabbit normal liv