摘要
目的 探讨迭代重组（IRIS）算法在低剂量肝脏增强CT扫描中的可行性及应用价值。方法 选择100例（男65例,女35例,平均年龄48岁）拟行肝脏三期增强的患者,常规剂量（120 k V）A组58例,采用滤波反投影算法（FBP）重组生成图像A;低剂量（100 k V）B组42例,分别采用FBP生成图像（B1组）及采用IRIS生成图像（B2组）,记录A与B两组的剂量长度乘积（DLP）、CT剂量容积指数（CTDI）并比较两组的有效剂量（ED）。计算正常肝脏门脉期的噪声及对比噪声比（CNR）,由3名有经验的放射科医师对血管边缘锐利度、图像主观噪声、病灶显示按1~5分评分。采用SPSS 20.0软件包对A、B两组的辐射剂量、图像质量进行统计学分析,检验水准α=0.05。结果 低剂量组ED[（1.23±0.32）m Sv]较常规剂量组ED[（2.30±0.73）m Sv]平均降低了46.24%。在3组图像客观噪声值中,B2组（8.85±1.11）较B1组（11.82±1.59）明显降低,不同算法的两组噪声值的差异有显著统计学差异（P〈0.001）,B2组与A组（8.09±1.52）算法无统计学意义（P〉0.05）;CNR分别为B2组（3.94±1.99）、B1组（3.46±2.38）、A组（3.87±2.19）,三组均无统计学差异（P〉0.05）;B2组图像各项诊断信息评分与A组无差异（P〉0.05）。结论 IRIS算法行肝脏增强扫描,可以降低图像噪声,提高CNR,在扫描剂量接近减半的情况下仍可获得满足诊断要求的肝脏增强CT图像。
Objective To investigate the feasibility of iterative reconstruction in image space （IRIS） algorithm in low- dose CT scanning of liver, and to discuss its clinical application. Methods One hundred patients （ 65 males and 35 fe- males, mean age of 48 years） , who were prepared to receive enhanced tri-phase CT scanning of liver, were included in this study. The patients were divided into group A （n = 58 ） and group B （n = 42）. In group A, routine dose （120kV） was employed and the raw data were reconstructed using the filtered back projection （FBP） algorithm to produce image A. In group B, low dose （100kV） was employed, and use FBP algorithm to produce image B1 and use IRIS algorithm to produce image B2. The dose-length product （DLP）, the CT volume index （CTVI） and the effective dose （ED） in both groups were determined. The normal liver noise in portal phase and the contrast-to-noise ratio （CNR） were calculated. The sharp- ness of vessels, the subjective image noise and the visualization of the lesion were evaluated by three qualified radiologists. Using SPSS20.0 software package the radiation dose, image quality of the two groups were statistically analyzed with the in- spection level of a = 0.05. Results The ED in group B was （ 1.23 ± 0.32） mSv, which was reduced by 46.24% when compared with that in group B （2.03 ±0.73 mSv）. The objective noise value of the group B2 and B1 was （8.85 ± 1.11 ） and （ 11.82 ± 1.59 ） respectively, the difference between the two groups was statistically significant （ P 〈 0.05 ）. The ob- jective noise value of the group A was （ 8.09 ± 1.52） , and the difference between the group A and group B2 was not signif- icant （P 〉 0.05）. CNR of group B2, B1 and A was （3.94 ± 1.99）, （3.46 ± 2.38） and （3.87 ± 2.19） respectively, the differences between each other among them were not significant （ P 〉 0.05 ）. The difference in evaluation scores between group B2 and group A was not significant （P 〉 0.05 ） . Conclusion Compared with standard filtered back projection al-gorithm, the iterative reconstruction in image space algorithm can reduce the image noise and increase CNR in enhanced CT scanning of liver; in the case of using approximately half of usual scan dose the images thus obtained can still meet the di- agnostic requirements.
出处
《临床放射学杂志》
CSCD
北大核心
2015年第1期123-127,共5页
Journal of Clinical Radiology
基金
国家自然科学基金资助项目(编号:81371661)
