AIDR3D重建算法超低剂量640层CT结肠成像息肉检出能力的实验研究

Experimental study on polyp detectability of ultralow-dose 640-slice CT colonography with adaptive iterative dose reduction three-dimensional reconstruction

目的实验研究三维自适应迭代剂量降低（AIDR3D）重建算法超低剂量640层CT结肠成像的息肉检出能力。方法制作10段猪结肠息肉模型,每段模型上有30枚1~15mm模拟息肉。所有模型用640层CT机扫描,球管电压120kVp,电流剂量分别为10、20、30、40、50mAs。滤波反投影（FBP）和AIDR3D算法重建后获得CT结肠成像的图像。两位不知道扫描电流剂量及重建算法的影像科医生独立分析图像。每位影像科医生的息肉检出信心分5个等级,用Kappa分析评价2位影像科医生息肉检出信心的一致性。根据剂量长度乘积计算辐射剂量。统计分析5种扫描电流剂量、2种重建算法的图像检出息肉的敏感性。结果两位影像科医生对3 000枚息肉检出信心一致性好（κ=0.740）。20mAs扫描并AIDR3D重建对1~〈6、6~〈11、11~15mm息肉检出敏感性与50mAs扫描并FBP重建相同,均达100%;有效辐射剂量1.62mSv较50mAs扫描并FBP重建的4.04mSv降低60.00%,差异有统计学意义（P〈0.05）。10mAs扫描并AIDR3D重建对1~〈6mm息肉检出敏感性77%低于50mAs扫描并FBP重建的100%,差异有统计学意义（P〈0.05）,对6~〈11mm息肉检出敏感性98%也低于50mAs扫描并FBP重建的100%。结论超低剂量20mAs扫描并AIDR3D重建640层CT结肠成像息肉检出能力能够与标准低剂量50mAs扫描并FBP重建媲美,检出敏感性达100%;而且有效辐射剂量降低60.0%,只有1.62mSv。

Objective To experimentally investigate polyp detectability of ultralow-dose 640-slice CT colonography with adaptive iterative dose reduction three-dimensional （AIDR3D） reconstruction. Methods Ten segments of porcine colonic phantom were established. There were thirty pedunculated polyps with diameters ranging between 1 mm and 15 mm on colon mucosal sur- faces of each segment. All phantoms were scanned using a 640-slice CT scanner with tube voltage of 120 kVp,and current doses of 10,20,30,40,50 mAs, respectively. CT colonography images were obtained after reconstructing scanning image data with filtered back projection （FBP） and AIDR3D algorithms. Two radiologists without knowing scanning current doses and reconstruction algo- rithms independently analyzed images. Confidence of polyp detection was divided into five grades for each radiologist, and its agree- ment was evaluated using Kappa analysis. Radiation dose was calculated based on dose-length product that was exported on CT scanner. Polyp detection sensitivities with five scanning current doses and two reconstruction algorithms were statistically analyzed. Results Agreement in confidence of 3 000 polyps detection by two radiologists was good （κ=0. 740）. Compared with 50mAs with FBP, 20 mAs with AIDR3D had the same detection sensitivity of 100 % for simulated polyps of 1- 〈6,6-〈 11 and.11- 15 mm. Effective radiation dose of 20 mAs with AIDR3D （1.62 mSv） was decreased by 60.00% compared with 50 mAs with FBP （4.04 mSv） ,their difference was statistically significant（P〈0.05）. As for 10 mAs with AIDR3D, detection sensitivity for simulated polyp of 1 -〈 6 mm （77 %） was lower than that using 50 mAs with FBP （ 100%）, their difference had statistically significance （ P 〈0.05） ;detection sensitivity for simulated polyp of 6-〈11 mm （98%） was also lower than that using 50 mAs with FBP（100%）. Conclusion Polyp detectability of ultralow-dose of 20 mAs 640-slice CT colonography with AIDR3D could comparable to that of standard low-dose of 50 mAs with FBP, with detection sensitivity of 100%,its effective radiation dose reduced by 60.0% and was only 1.62 mSv.