能谱CT虚拟平扫联合多模型迭代重建算法在甲状腺图像质量及辐射剂量中的应用

Adaptive Statistical Iterative Reconstruction-V Algorithm Combined with Virtual Non-Contrast Scans in Thyroid Image Quality and Radiation Dose

目的探讨能谱CT虚拟平扫(VNC)联合多模型迭代重建算法(ASiR-V)在甲状腺图像质量及辐射剂量应用中的可行性。资料与方法收集需进行甲状腺增强检查的25例患者,采取非能谱平扫及能谱增强扫描,检查结束后在静脉期重建不同权重ASiR-V(0、20%、40%、60%、80%、100%),并重建不同权重ASiR-V的VNC,分别记作A1~A6。平扫按照滤波反投影重建,记作B1;记录各期检查辐射剂量,测量并计算6组图像感兴趣区的CT值、图像噪声(SD)及信噪比(SNR),并对图像主观和客观指标进行分析。结果A1、B1两组图像中背景噪声SD值(SD脂肪)、SNR比较,差异均无统计学意义(t=−0.196、−0.778,P均>0.05),B1图像中甲状腺CT值、SD值均高于A1图像(t=−9.206、−6.970,P均<0.05)。A1~A6图像中甲状腺SD值、SNR比较,差异均有统计学意义(F=63.258、3.987,P<0.05),随着迭代强度增大,图像SD值逐渐减小(R2=0.598,P<0.05);A1~A6图像中甲状腺CT值差异无统计学意义(F=0.130,P>0.05);A1~A6图像的0~60%主观评分呈上升趋势,40%~60%主观评分最佳(Kappa=0.83,P<0.05),80%~100%图像评分呈下降趋势,与60%图像比较差异有统计学意义(Z=−3.60、−5.47,P<0.05),权重60%的VNC主观评分高于真实平扫(Z=−4.171,P<0.05)。应用能谱虚拟平扫代替常规平扫后25例检查者的有效剂量可降低47.9%,其中静脉期能谱扫描与非能谱的有效剂量分别为(1.66±0.23)mSv、(1.75±0.24)mSv,差异有统计学意义(t=5.261,P<0.05)。结论甲状腺静脉期VNC可应用于临床诊断,降低辐射剂量47.9%;且联合后置60%权重能明显降低图像噪声,改善图像质量。

Purpose To explore the feasibility of adaptive statistical iterative reconstruction-V(ASiR-V)combined with virtual non-contrast scan(VNC)in thyroid image quality and radiation dose in the clinical application.Materials and Methods Twenty-five patients prepared for CT thyroid enhanced scan were prospectively collected.All patients received conventional true non-contrast(TNC)scanning and spectral enhanced scanning.The venous phase used spectral mode scans was used to reconstruct VNC images with different ASiR-V levels(0,20%,40%,60%,80%,100%),which were recorded as A1-A6 group,respectively.Plain scan was reconstructed by filtered back projection and recorded as B1.Radiation dosage of every phase was recorded,and the CT value,image noise(SD)and signal noise ratio(SNR)of region of interest of six groups were measured and compared,respectively.Results There was no significant difference in image background noise(SDfatty)and SNR between A1 group and B1 group(t=−0.196 and−0.778,all P>0.05).CT thyroid value and SD thyroid of B1 group image were significant higher than that of A1 group(t=−9.206,−6.970,all P<0.05).There were all statistically significant differences in image noise and SNR among A1-A6 groups(F=63.258,3.987,all P<0.05).With the increase of ASiR-V level,the image noise gradually decreased(R2=0.598,P<0.05).There was no significant difference in CT thyroid value among A1-A6 groups.With the increase of ASiR-V level,the subjective scores of images showed gradually increases.Among A1-A6 groups,the subjective score of 40%-60%ASiR-V image was the best(Kappa=0.83,P<0.05).The subjective score of 80%-100%ASiR-V image showed a downward trend,and the difference was statistically significant compared with that of 60%ASiR-V image(Z=−3.60,−5.47,P<0.05).The subjective score of VNC image with 60%ASiR-V image was significantly higher than that of TNC image.The effective dose(ED)of 25 patients was decreased by 47.9%after using energy spectrum virtual scanning replacing conventional scanning.The ED of venous energy spectrum scanning and non-energy spectrum scanning were(1.66±0.23)mSv and(1.75±0.24)mSv,respectively,and there was significant difference in the ED among A1-A6 groups(t=5.261,P<0.05).Conclusion VNC scan of thyroid venous phase can meet the requirements of imaging diagnosis,reducing radiation dose by 47.9%.VNC scan combined with 60%ASiR-V in the vein phase of thyroid energy spectrum scan can significantly reduce radiation dose and improve image quality in clinical application.