全模型迭代重建算法在头颈CT血管成像中改善图像质量的应用

Application of Forward Projected Model-Based Iterative Reconstruction Solution in Improving Image Quality of Head and Neck CT Angiography

目的探讨全模型迭代重建(FIRST)算法对100 kV低剂量头颈CT血管成像(CTA)图像质量改善的价值。资料与方法回顾性收集2021年9—10月在北京医院行头颈CTA检查的患者32例,对其原始数据进行重建并分为滤波反投影(FBP)算法图像组、混合迭代(AIDR 3D)算法图像组及FIRST算法图像组。比较3组图像头颈部主要血管(颈总动脉,椎动脉,颈内动脉C1、C4段,大脑中动脉M1、M3段,大脑前动脉A1、A3段)、背景(肌肉、脑组织)的平均CT值、噪声、信噪比及各血管的对比噪声比,并比较3组图像质量的主观评分。以数字减影血管造影(DSA)为“金标准”,评估各重建算法对头颈部血管狭窄的检出情况。结果与AIDR 3D组、FBP组相比,FIRST组显著降低了头颈部血管以及背景(肌肉、脑组织)的图像噪声(t=-13.19~-7.28,P均<0.001),其信噪比、对比噪声比明显高于FBP组与AIDR 3D组(t=2.17~9.67,P均<0.001)。颈总动脉、颈内动脉C1、C4段,大脑中动脉M1段、大脑前动脉A1段FIRST组的CT值高于AIDR 3D组(t=1.28~3.60,P均<0.05);FIRST组背景(肌肉、脑组织)的CT值显著低于FBP组与AIDR 3D组(t=-7.63~-4.03,P均<0.001),差异有统计学意义。对于头颈部血管的显示,FIRST组与AIDR 3D组图像满足诊断要求,主观评分均高于FBP组图像(P均<0.05),且FIRST组图像评分高于AIDR 3D组(P<0.05)。与DSA相比,FBP算法对于轻、中度狭窄诊断效能不佳,FIRST、AIDR 3D算法对血管狭窄程度评估结果与DSA一致。结论在100 kV低剂量头颈CTA联合扫描中,与传统FBP、AIDR 3D算法相比,FIRST算法可以有效改善头颈CTA图像质量,满足头颈部血管疾病的诊断要求。

Purpose To explore the application value of forward projected model-based iterative reconstruction solution(FIRST)algorithm improving image quality of 100 kV low-dose head and neck CT angiography(CTA).Materials and Methods A total of 32 patients who underwent head and neck CTA examinations in the Beijing Hospital from September to October 2021 were retrospectively recruited.All patients were divided into three groups based on different reconstruction,including the filtered back projection(FBP)group,the adaptive iterative dose reduction 3D(AIDR 3D)group,and the FIRST group.The average CT value,image noise,signal-to-noise ratio,contrast to noise ratio and other objective indicators of the main blood vessels images(including common carotid artery,vertebral artery,internal carotid artery C1,C4,middle cerebral artery M1,M3 and anterior cerebral artery A1,A3)and background(including muscle,brain tissue)in the head and neck of the three groups were recorded and analyzed,respectively.The subjective imaging evaluation was scored and analyzed.Digital subtraction angiography(DSA)was used as the gold standard to evaluate the detection of vascular stenosis in the head and neck by each reconstruction algorithm.Results Compared with the AIDR 3D group and the FBP group,the FIRST group significantly reduced the image noise of the blood vessels and the background(muscle,brain tissue)(t=-13.19--7.28,all P<0.001).The signal-to-noise ratio and contrast to noise ratio of the FIRST group were significantly higher than those of the FBP group,AIDR 3D group(t=2.17-9.67,all P<0.001).The CT values of FIRST group were significantly higher than those of AIDR 3D group in the common carotid artery,internal carotid artery C1 and C4,middle cerebral artery M1,and anterior cerebral artery A1 segment(t=1.28-3.60,all P<0.05).The CT value of background(muscle,brain tissue)in FIRST group was significantly lower than those in the FBP group and the AIDR 3D group,with statistically significant difference(t=-7.63--4.03,all P<0.001).The images of FIRST group and AIDR 3D group met the diagnostic requirements,and the subjective scores of the two groups were significantly higher than those of the FBP group images(all P<0.05).The imaging scores of FIRST group were significantly higher than those of AIDR 3D group(P<0.05).Compared with DSA,FBP was ineffective for mild and moderate stenosis;FIRST and AIDR 3D algorithms were consistent with DSA.Conclusion Compared with traditional FBP and AIDR 3D reconstruction algorithms,FIRST algorithm can effectively improve the image quality of low-dose head and neck CTA and obtain better image quality and meet the diagnostic performance of head and neck vascular diseases.