摘要
目的:探讨人工智能(AI)成像优化技术联合迭代算法(ClearView+)对"双低"(低辐射剂量及低碘负荷)主动脉CTA图像质量的影响。方法:前瞻性连续纳入2018年2～5月在本院NeuViz128CT行主动脉CTA检查的40例患者,随机分为A、B两组,每组20例。A组为"双低"组(管电压80kVp,对比剂40mL);B组为常规组(管电压120kVp,对比剂80mL)。A组图像采用ClearView+迭代算法分别按照6个不同重建档位(0%、10%、30%、50%、70%及90%;0%为滤波反投影)重建图像,选出图像质量最佳重建档位作为A1组,进一步对A1组图像采用AI成像优化技术进行图像优化,所得图像作为A2组;B组则采用滤波反投影重建算法。对A1,A2和B 3组图像进行客观评价,分别以测量3组主动脉感兴趣区的CT值及噪声(SD)值,计算信噪比(SNR)及对比噪声比(CNR)作为图像质量客观评价指标,主观评价由2名高年资放射科医生采用双盲法进行5分制评价(5分-优良,1分-极差)。记录和比较A组与B组的对比剂用量和辐射剂量(CTDIvol,DLP和ED)。结果:ClearView+90%档位分别与0%、10%、30%、50%和70%档位比较,主动脉SD显著降低、SNR及CNR显著升高(P<0.05),因此Clearview+90%作为最佳图像质量的迭代档位。三组图像定量评价,SD、SNR、CNR两两比较结果:A2组SD(12.21±2.66)<A1组SD(17.67±2.75)<B组SD(21.83±4.64)(P<0.05),A2组SNR(36.77±10.13)>A1组SNR(29.57±7.41)>B组SNR(19.09±5.65)(P<0.05),A2组CNR(47.92±12.36)>A1组CNR(38.50±8.96)>B组CNR(23.37±6.61)(P<0.05)。三组图像之间主观评分无统计学差异(P>0.05)。"双低"组与常规组的辐射剂量评价指标CTDIvol(2.82±0.36vs 13.34±1.84)mGy,DLP(181.21±21.98vs 871.17±134.26)mGy·cm和ED(2.54±0.31vs 12.20±1.88)mSv均有显著性差异。"双低"组有效辐射剂量ED较常规组显著降低79.18%,对比剂用量降低50%。结论:AI成像优化技术联合ClearView+迭代算法,使NeuViz 128CT主动脉"双低"扫描可以获得和常规扫描同样的图像质量且有效地降低了辐射剂量和对比剂用量。
Objective:To investigate the value of artificial intelligence(AI)-based image optimization technique and iterative algorithm(ClearView+)in the image quality of"double low"aorta CTA.Methods:Forty patients who were referred to perform aorta CTA examination between February and May 2018 in NeuViz128 CT of our hospital were prospectively collected and randomly divided into two groups.Group A was scanned using a low tube voltage of 80kVp and a reduced contrast medium volume of 40mL ioversol(320mg I/mL),and group B was scanned with a standard 120kVp tube voltage and an injection of 80mL contrast.Group A was reconstructed using ClearView+technique of six percentages including 0%(filtered back projection),10%,30%,50%,70%and 90%.After comparing these different percentages,the best percentage level of ClearView+was selected as group A1.Both best percentage of iterative algorithm and further AI-based image optimization were used in group A2.Group B was reconstructed by using filtered back projection(FBP).The mean attenuation of contrast-enhancement values,image noise,signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR)were calculated to evaluate image quality objectively.In addition,subjective image analysis was performed using five-point scales independently by two radiologists(5-excellent,1-undiagnostic).The estimated radiation dose in terms of CT dose index(CTDIvol),dose length product(DLP),and effective dose(ED)was recorded and compared between group A and group B.Results:Compared with other percentage of ClearView+,the image noise of Clearview+90%was decreased and whereas SNR and CNR increased significantly(P<0.05).Therefore,90%was the best percentage level of Clearview+without over-smoothing.The comparison results of SD,SNR and CNR among the three groups were below:group A2 SD(12.21±2.66)<group A1SD(17.67±2.75)<group B SD(21.83±4.64)(P<0.05),group A2 SNR(36.77±10.13)>group A1 SNR(29.57±7.41)>group B SNR(19.09±5.65)(P<0.05),A2 group CNR(47.92±12.36)A1 group CNR(38.50±8.96)>B group CNR(23.37±6.61)(P<0.05).No statistical differences were observed among the three groups in the scores of subjective image quality(P>0.05).There were significant differences between the"double low"group and the conventional group in the evaluation indexes of radiation dose including CTDIvol(2.82±0.36 vs 13.34±1.84)mGy,DLP(181.21±21.98 vs 871.17±134.26)mGy·cm and ED(2.54±0.31 vs 12.20±1.88)mSv.Compared with the conventional group,the effective dose and contrast medium volume of the"double low"group were reduced by 79.18%and 50%respectively.Conclusions:With the combination of ClearView+and AI-based image optimization technique,the"double low"aorta CTA scan of NeuViz 128 CT could provide the consistent image quality with the regular scan,while magnificently reducing the radiation dose and contrast medium intake.
作者
王明
王怡宁
于敏
王沄
王曼
金征宇
WANG Ming;WANG Yining;YU Min(Department of Radiology,Peking Union Medical College Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College,Beijing 100730,China)
出处
《放射学实践》
北大核心
2018年第10期1009-1016,共8页
Radiologic Practice
