多层螺旋CT下肢动脉血管成像:容积再现及最大强度投影图像质量的影响因素

Multi-detector CT Angiography of Aortoiliac and Lower Extremity Arteries: Effect Factors on Image Quality of Volume Rendering and Maximum Intensity Projection

目的分析容积再现(VR)、最大强度投影(MIP)图像质量与图像噪声(N)、信噪比(SNR)和对比噪声比(CNR)的相关性。方法回顾性分析2017年6-11月临床拟诊周围动脉闭塞性疾病(PAOD)或术前需进行下肢动脉CT血管成像(CTA)检查的61例患者的临床资料。患者采用东芝Aquilion 64层螺旋CT机扫描成像,分别以卷积核FC11,FC12,FC13,FC14,FC15重建主-髂段图像,将数据传至工作站重建VR和MIP图像,并测量主-髂段血管增强CT值、对比度、N、背景信号、SNR和CNR;VR和MIP主观图像质量由2名医师采用盲法独立完成,最后结果以取得一致意见为准。统计采用多变量相关性分析、t检验。结果以VR和MIP图像等级评价为标准,采用Spearman秩相关分析CNR、SNR及N与图像质量的相关性,结果发现,与VR图像相关性:CNR=0.73,SNR=0.72,N=-0.59;与MIP图像相关性:CNR=0.61,SNR=0.60,N=-0.48。优秀的MIP图像与优秀的VR图像N、SNR及CNR 3个参数比较,差异有统计学意义(N:t=3.419,P=0.001;SNR:t=-3.397,P=0.001;CNR:t=-3.467,P=0.001),VR对源图像质量的要求更高。以FC11为标准处理,其MIP和VR图像质量与血管增强CT值、对比度、SNR、CNR及N的相关,与N相关性最小。结论与VR和MIP图像质量相关性最高的是CNR,其次为SNR,CTA三维图像质量主要与血管强化值有关,噪声的影响较小。VR对源图像质量的要求更高。

Objective To analyse the correlation between imaging parameters,such as noise(N), signal-to-noise ratio(SNR), contrast-to-noise ratio(CNR), and image quality of volume rendering(VR) and maximum intensity projection(MIP). Methods The clinical data of 61 patients who were clinically diagnosed with peripheral artery occlusion disease(PAOD) or who needed to undergo CT angiography(CTA) examination of the lower limb artery from June to November 2017 were retrospectively analyzed. The original axial images were acquired with the equipment of aquilion 64-MDCT of Toshiba, and five retrospectively reconstructed algorithm(FC11,FC12,FC13,FC14,FC15). Vitrea workstation was used for further three-dimensional reconstruction, such as MIP and VR. CT value of aorta, N and background signal were recorded;contrast, SNR, CNR were calculated. Two experienced readers evaluated the image quality of the axial sections as well as 3D reconstructions image with blind method. Multivariate correlation analysis and t test were used. Results Spearman rank correlation was used to analyze the correlation between CNR, SNR and N and image quality. The results showed that CNR = 0.73, SNR = 0.72 and N=-0.59 were correlated with VR images.Correlation with MIP images: CNR = 0.61, SNR = 0.60, N =-0.48.There were statistically significant differences between excellent MIP images and excellent VR images in N, SNR and CNR parameters(N: t=3.419, P=0.001;SNR: t=-3.397, P=0.001;CNR: t=-3.467, P=0.001), VR required higher quality of source image.With FC11 as the standard, the MIP and VR image quality were correlated with the value of vascular enhanced CT, contrast, SNR, CNR and N, with the least correlation with N. Conclusion CNR has the highest correlation with VR and MIP image quality, followed by SNR. The quality of CTA 3D image is mainly related to the value of vascular enhancement, with little impact of noise.VR requires higher quality of source images.