超低剂量条件下IMR技术对肺容积定量评估及胸部CT图像质量的影响  

Ultralow Dose Chest CT:The Influence of Quantitative Assessment in Lung Volume and Image Quality Using Iterative Model Reconstruction

作　　者：方元[1] 管宇[1] 夏艺[1] 范丽[1] 刘士远[1] 肖湘生[1] FANG Yuan;GUAN Yu;XIA Yi(Department of Radiology, Changzheng Hospital of the Second Military Medical university, Shanghai 20003, P. R. Chin)

机构地区：[1]第二军医大学附属长征医院影像科

出　　处：《临床放射学杂志》2018年第5期860-864,共5页Journal of Clinical Radiology

摘　　要：目的探讨在超低剂量条件下全模型迭代重组技术（IMR）对肺容积定量评估及胸部CT图像质量的影响。方法搜集31例行胸部CT检查的体检者,分别用低剂量（120 k Vp,50 m As）和超低剂量（120 k Vp,10 m As）条件进行扫描,低剂量组（LD组）用滤波反投影法（FBP）进行重组,超低剂量组（ULD组）分别用FBP、高级迭代重组算法（i Dose4）以及IMR的Routine和Soft Tissue算法（均为等级1）进行重组,对重组后图像进行肺容积定量参数和客观噪声值的测量,并用5分法对不同重组方式的主观图像质量进行主观评分。用统计学方法比较各组定量指标、客观图像噪声及主观评分间的差异。结果超低剂量条件下,和LD组相比,FBP容积定量测量的准确性受到较大影响,IMR的Soft Tissue算法也受到一定影响,i Dose4和Routine算法均能保证定量结果的准确性。客观噪声方面,FBP-ULD组的客观噪声最高,和其相比,i Dose4-ULD组、IMR-R1-ULD组和IMR-ST1-ULD组噪声分别减少29.3%、55.4%和70.1%（P均〈0.05）,而和i Dose4-ULD组相比,IMR-R1-ULD组和IMR-ST1-ULD组噪声分别减少36.9%和57.7%（P均〈0.05）。FBP-LD组的主观图像质量最高,其次为IMR-R1-ULD组和IMR-ST1-ULD组,3组间均无明显统计学差异（P〉0.05）;FBP-ULD组最低,和其余组别均存在明显统计学差异（P均=0.000）;i Dose4-ULD组高于FBP-ULD但低于FBP-LD和IMR-R1-ULD两组（P均〈0.05）,但和IMR-ST1-ULD组无明显差异（P=0.220）。斑污现象只存在IMR组别中,Soft Tissue组较Routine组明显,且组间存在差异（P=0.000）,斑污现象并不明显影响图像的整体质量。结论在超低剂量条件下,推荐使用IMR的Routine算法进行肺容积定量分析,既能保证定量结果的准确,又能很好地降低图像噪声,提高图像质量。Objective To investigate the influence of quantitative assessment in lung volume and image quality under ultralow radiation dose condition by using iterative model reconstruction（IMR）.Methods 31 patients were prospectively enrolled and underwent scanning twice with low-dose CT（reference parameters,120 k Vp,50 m As） and ultralow-dose CT（reference parameters,120 k Vp,10 m As）.The original data of low-dose group was reconstructed by FBP（contrast group）,and the ultralow-dose group were reconstructed by FBP,i Dose4 and IMR respectively.Calculated the lung volume quantitative parameters and objective image noise（OIN） for each group,and assessed the image quality using 5 point scale（poor to excellent）,then compared the quantitative and qualitative parameters among all groups.Results Under ultra-low dose conditions,compared with the control group,the accuracy of the quantitative measurement by the FBP method was greatly affected.The IMR Soft Tissue algorithm was also affected.The i Dose4 and Routine algorithms both ensure the accuracy of the quantitative results.In terms of objective noise,the objective noise of the FBP-ULD group was the highest,compared with the noise of the i Dose4-ULD group,the IMR-R1-ULD group,and the IMR-ST1-ULD group,which decreased by29.3%,55.4%,and 70.1%,respectively（P〈0.05）,compared with i Dose4-ULD group,the noise of IMR-R1-ULD group and IMR-ST1-ULD group was reduced by 36.9% and 57.7%,respectively（P〈0.05）.The subjective image quality of the control group（FBP-LD group） was the highest,followed by the IMR-R1-ULD group and the IMR-ST1-ULD group.There was no statistically significant difference among the three groups,and the FBP-ULD group was the lowest,and the rest of the group.There were significant statistical differences（P = 0.000）,i Dose4-ULD group was higher than FBP-ULD but lower than FBP-LD and IMR-R1-ULD groups（P〈0.05）,but with IMR-ST1-ULD.There was no significant difference between the groups（P = 0.220）.The spotting phenomenon was onl

关 键 词：计算机体层成像 全模型迭代重组 定量评估 超低剂量 肺部 

分 类 号：R445.2[医药卫生—影像医学与核医学] R816.4[医药卫生—诊断学]