AIDR 3D技术在肝脏低辐射剂量和低对比剂用量增强CT中的应用研究
发布时间：2018-08-13 17:09
【摘要】：第一部分水模实验目的:检测三维自适应迭代剂量降低(Adaptive Iterative Dose Reduction 3D,AIDR 3D)算法的降噪能力。评估不同管电压对图像噪声的影响程度。材料和方法:采用普通水模(直径25cm),按照120KV,滤波反投影(Filtered Back Projection,FBP)重建;120KV,AIDR 3D重建;100KV,AIDR 3D重建;80KV,AIDR 3D重建四组扫描方案,以不同的噪声指数(Noise Index,NI)(NI 5-11,间隔0.5)对水模进行扫描,测量四组图像的噪声,测算AIDR 3D的降噪能力。结果:采用120KV+AIDR 3D重建算法图像噪声比120KV+FBP重建图像噪声降低,两者间具有统计学差异(q=7.131,P0.001);100KV+AIDR 3D重建算法图像噪声比120KV+FBP重建算法图像噪声降低,两者间具有统计学意义(q=6.064,P0.001);80KV+AIDR 3D重建算法比100KV+AIDR 3D重建算法图像噪声增加,两者间具有统计学差异(q=3.888,P0.05)。结论:①与FBP算法相比,AIDR 3D重建算法能明显降低图像噪声。②管电压为80KV时图像噪声较100KV及120KV时明显增高。第二部分临床实验目的:探讨自动管电流调节技术和低管电压结合自动管电流调节技术联合三维自适应迭代剂量降低(Adaptive Iterative Dose Reduction 3D,AIDR 3D)算法在低对比剂剂量肝脏增强CT中的应用价值。材料和方法:前瞻性的将150例行常规肝脏增强CT的患者按随机表分为3组(A,B,C),每组50例,A组为常规组,采用FBP重建+常规对比剂用量(1.5ml/Kg),B组及C组为双低组,二者管电压不同,其中B组为120KV,C组为100KV,均采用AIDR 3D重建+低对比剂用量(1.0ml/Kg)。记录每组图像的CT容积剂量指数(CT Dose index-volume,CTDI vol)、剂量长度乘积(Dose Length Product,DLP),并计算出3组的有效剂量(Effective Dose,ED),平均CT值、图像噪声、信噪比(SNR)、对比信噪比(CNR)。对3组图像的诊断信息(图像主观噪声、图像总体质量)按1-4分(1分最差,4分最优)予以评分。计量资料采用完全随机设计的方差分析及秩和检验。计数资料采用完全随机设计的非参数检验(Kruskal-Wallis)对结果进行比较。结果:双低组(B组及C组)的有效剂量较常规组(A组)减低(A、B、C组分别为2.98±1.33,2.23±0.75,2.54±0.55),A组与B组间差异具有统计学差异(F=8.10,t=4.004,P=0.000,0.01),A组与C组间差异亦具有统计学意义(F=8.10,t=2.348,P=0.020,0.05);B组和C组之间差异无统计学意义(p0.05)。图像质量客观评估中,A组图像肝实质、主动脉及门静脉噪声最高,C组最低,三组之间差异具有统计学意义(肝实质:F=216.06,主动脉:F=150.83,门静脉:F=150.61;P均等于0.000,0.01)。A组与C组图像肝实质、主动脉及门静脉CT值无统计学差异(p0.05),B组最低,与A组和C组间有统计学差异(肝实质:F=38.79,主动脉:F=52.78,门静脉:F=56.19,P均等于0.000,P0.01)。B组的CNR与A组相比,无统计学差异(p0.05);C组的CNR高于A组和B组(A组VS C组:F=37.62,t=7.010,P=0.000,0.01;B组VS C组:F=37.62,t=7.937,P=0.000,0.01)。C组的SNR最高,A组最低,三组间差异具有统计学意义(A组VS B组:F=162.36,t=3.096,P=0.000,0.01;A组VS C组:F=162.36,t=16.936,P=0.000,0.01;B组VS C组:F=162.36,t=13.84,P=0.000,0.01)。主观评估中,B组和C组的图像质量评分均高于A组,存在显著统计学差异(A组VS B组:H=-5.288,P=0.000,0.01;A组VS C组:H=-5.688,P=0.000,0.01)。C组的图像质量评分高于B组,差异无统计学意义(P0.05)。结论:①AIDR 3D重建结合低对比剂用量较FBP重建结合常规对比剂用量在腹部CT增强中能得到相当甚至更好的图像质量;②低管电压联合自动管电流调节技术较单纯自动管电流调节在肝脏低对比剂用量增强CT中能得到更好的图像质量。
[Abstract]:The first part of the water model experiment is to test the noise reduction ability of the adaptive Iterative Dose Reduction 3D (AIDR 3D) algorithm and evaluate the influence of different tube voltages on image noise. Build; 120KV, AIDR 3D reconstruction; 100KV, AIDR 3D reconstruction; 80KV, AIDR 3D reconstruction four groups of scanning scheme, with different noise index (NI) (NI 5-11, interval 0.5) to scan the water model, measure the noise of the four groups of images, calculate the noise reduction ability of AIDR 3D. Results: 120KV + AIDR 3D reconstruction algorithm image noise ratio 120KV + FBP reconstruction image noise reconstruction. The image noise of 100KV+AIDR 3D reconstruction algorithm was lower than that of 120KV+FBP reconstruction algorithm (q=6.064, P 0.001), and that of 80KV+AIDR 3D reconstruction algorithm was higher than that of 100KV+AIDR 3D reconstruction algorithm (q=3.888, P 0.05). Conclusion: Compared with FBP algorithm, AIDR 3D reconstruction algorithm can significantly reduce image noise. 2) The image noise at 80 KV tube voltage is significantly higher than that at 100 KV and 120 KV tube voltage. Materials and Methods: 150 patients with routine hepatic contrast-enhanced CT were prospectively divided into three groups (A, B, C) according to the randomized table, 50 cases in each group, 50 cases in group A, FBP reconstruction + routine contrast medium dosage (1.5ml/Kg), and two low groups in group B and C. The CT Dose Index-volume (CTDI vol), Dose Length Product (DLP) and effective dose (ED), mean CT value, image noise were recorded for each group. The diagnostic information (subjective noise, overall image quality) of the three groups of images was scored by 1-4 points (the worst one, the best four). The measurement data were analyzed by variance analysis and rank sum test. The counting data were analyzed by Kruskal-Wallis. Results: The effective dose of double-low group (group B and group C) was lower than that of group A (group A, group B and group C were 2.98 [1.33, 2.23] 0.75, 2.54 [0.55] respectively). There were significant differences between group A and group B (F = 8.10, t = 4.004, P = 0.000, 0.01). There were also significant differences between group A and group C (F = 8.10, t = 2.348, P = 0.020, 0.05). In the objective evaluation of image quality, the liver parenchyma, aorta and portal vein noise were the highest in group A, and the lowest in group C. There were significant differences among the three groups (liver parenchyma: F = 216.06, aorta: F = 150.83, portal vein: F = 150.61; P = 0.000, 0.01). There was no unified CT value between group A and group C. The difference was statistically significant (p0.05), the lowest in group B, and the lowest in group A and C (liver parenchy: F = 38.79, ao: F = 52.78, portal ve: F = 56.19, P = 56.19, P = 0.000, P = 0.01). There was no significant difference in CNR between group B and group A (p0.05); the CNR in group C was higher in group C than group A and group B (VC group A: F = 37.62, t = 37.62, t = 7.62, t = 7.010, t = 7.010.01, P = 0.01; VC group B: F = 37.62, F = 37.62, P = 37.62, P = 7.62, t = 7.62, t = 7.937, t In the meantime, it is necessary to study the relationship between the two. The SNR of group C was the highest, and that of group A was the lowest (group A vs group B: F = 162.36, t = 3.096, P = 0.000, 0.01); group A vs group C: F = 162.36, t = 16.936, P = 0.000, 0.01; group B vs group B: F = 162.36, t = 13.84, P = 0.000, 0.01). Group H = - 5.288, P = 0.000, 0.01; Group A: VS C: H = - 5.688, P = 0.000, 0.01). The image quality score of group C was higher than that of group B, and there was no significant difference (P 0.05). Conclusion: The image quality of AIDR 3D reconstruction combined with low contrast agent dosage was better than that of FBP reconstruction combined with conventional contrast agent dosage in abdominal CT enhancement. Automated tube current regulation technique can obtain better image quality than auto tube current regulation in contrast enhanced CT with low dose of contrast medium.
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R816.5
[Abstract]:The first part of the water model experiment is to test the noise reduction ability of the adaptive Iterative Dose Reduction 3D (AIDR 3D) algorithm and evaluate the influence of different tube voltages on image noise. Build; 120KV, AIDR 3D reconstruction; 100KV, AIDR 3D reconstruction; 80KV, AIDR 3D reconstruction four groups of scanning scheme, with different noise index (NI) (NI 5-11, interval 0.5) to scan the water model, measure the noise of the four groups of images, calculate the noise reduction ability of AIDR 3D. Results: 120KV + AIDR 3D reconstruction algorithm image noise ratio 120KV + FBP reconstruction image noise reconstruction. The image noise of 100KV+AIDR 3D reconstruction algorithm was lower than that of 120KV+FBP reconstruction algorithm (q=6.064, P 0.001), and that of 80KV+AIDR 3D reconstruction algorithm was higher than that of 100KV+AIDR 3D reconstruction algorithm (q=3.888, P 0.05). Conclusion: Compared with FBP algorithm, AIDR 3D reconstruction algorithm can significantly reduce image noise. 2) The image noise at 80 KV tube voltage is significantly higher than that at 100 KV and 120 KV tube voltage. Materials and Methods: 150 patients with routine hepatic contrast-enhanced CT were prospectively divided into three groups (A, B, C) according to the randomized table, 50 cases in each group, 50 cases in group A, FBP reconstruction + routine contrast medium dosage (1.5ml/Kg), and two low groups in group B and C. The CT Dose Index-volume (CTDI vol), Dose Length Product (DLP) and effective dose (ED), mean CT value, image noise were recorded for each group. The diagnostic information (subjective noise, overall image quality) of the three groups of images was scored by 1-4 points (the worst one, the best four). The measurement data were analyzed by variance analysis and rank sum test. The counting data were analyzed by Kruskal-Wallis. Results: The effective dose of double-low group (group B and group C) was lower than that of group A (group A, group B and group C were 2.98 [1.33, 2.23] 0.75, 2.54 [0.55] respectively). There were significant differences between group A and group B (F = 8.10, t = 4.004, P = 0.000, 0.01). There were also significant differences between group A and group C (F = 8.10, t = 2.348, P = 0.020, 0.05). In the objective evaluation of image quality, the liver parenchyma, aorta and portal vein noise were the highest in group A, and the lowest in group C. There were significant differences among the three groups (liver parenchyma: F = 216.06, aorta: F = 150.83, portal vein: F = 150.61; P = 0.000, 0.01). There was no unified CT value between group A and group C. The difference was statistically significant (p0.05), the lowest in group B, and the lowest in group A and C (liver parenchy: F = 38.79, ao: F = 52.78, portal ve: F = 56.19, P = 56.19, P = 0.000, P = 0.01). There was no significant difference in CNR between group B and group A (p0.05); the CNR in group C was higher in group C than group A and group B (VC group A: F = 37.62, t = 37.62, t = 7.62, t = 7.010, t = 7.010.01, P = 0.01; VC group B: F = 37.62, F = 37.62, P = 37.62, P = 7.62, t = 7.62, t = 7.937, t In the meantime, it is necessary to study the relationship between the two. The SNR of group C was the highest, and that of group A was the lowest (group A vs group B: F = 162.36, t = 3.096, P = 0.000, 0.01); group A vs group C: F = 162.36, t = 16.936, P = 0.000, 0.01; group B vs group B: F = 162.36, t = 13.84, P = 0.000, 0.01). Group H = - 5.288, P = 0.000, 0.01; Group A: VS C: H = - 5.688, P = 0.000, 0.01). The image quality score of group C was higher than that of group B, and there was no significant difference (P 0.05). Conclusion: The image quality of AIDR 3D reconstruction combined with low contrast agent dosage was better than that of FBP reconstruction combined with conventional contrast agent dosage in abdominal CT enhancement. Automated tube current regulation technique can obtain better image quality than auto tube current regulation in contrast enhanced CT with low dose of contrast medium.
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R816.5
【相似文献】
相关期刊论文 前10条
1 金燕南;赵俊;刘尊钢;庄天戈;;三源螺旋CT精确重建算法在动态重建中的应用[J];航天医学与医学工程;2009年06期
2 郑健;俞航;蒯多杰;刘兆邦;董月芳;张涛;;多层CT重建算法对比研究[J];CT理论与应用研究;2012年04期
3 白玫;杨雨;严汉民;;多排螺旋CT自适应统计迭代重建算法的客观评价[J];中国医学装备;2013年01期
4 程明渊;贺奇才;胡琴明;陈弟虎;;基于穿越长度权重迭代重建算法的研究[J];中国医学物理学杂志;2013年02期
5 闫镔;韩玉;魏峰;李磊;李建新;;锥束CT超视野成像重建算法综述[J];CT理论与应用研究;2013年02期
6 王小璞,张朋,李兴东,张兆田;一种块迭代的快速代数重建算法(英文)[J];CT理论与应用研究;2000年S1期
7 陈志强,李亮,康克军,张丽;锥束CT 精确重建算法研究最新进展(英文)[J];CT理论与应用研究;2005年03期
8 朱扬明;庄天戈;;有限角图象重建的虚拟对称重建算法[J];CT理论与应用研究;1992年01期
9 史颖琴;潘晋孝;;Feldkamp-type-VOI锥束CT重建算法加速的研究[J];CT理论与应用研究;2011年01期
10 吕东辉,庄天戈,严壮志;体积CT中的图象重建算法研究综述[J];CT理论与应用研究;2000年04期
相关会议论文 前10条
1 胡永胜;;重建算法对低剂量肺部CT扫描图像质量和诊断的影响[A];2010中华医学会影像技术分会第十八次全国学术大会论文集[C];2010年
2 伍晓平;谷士文;费耀平;刘应龙;李杰;;基于遗传算法的图象重建算法[A];中国图象图形科学技术新进展——第九届全国图象图形科技大会论文集[C];1998年
3 张朋;张兆田;;几种CT图像重建算法的比较[A];第九届中国体视学与图像分析学术会议论文集[C];2001年
4 宋e，
本文编号：2181642
本文链接：https://www.wllwen.com/yixuelunwen/yundongyixue/2181642.html
最近更新
教材专著
