在颞骨MSCT检查中应用迭代算法:实验研究

目的探讨在颞骨MSCT检查中应用迭代算法(采用iDose算法)对图像质量及辐射剂量的影响。方法以临床颞骨MSCT常规使用的滤波反投影算法(FBP)的噪声水平为基准,选择合适的离体头颅标本作为研究对象,改变有效管电流进行扫描,然后分别使用FBP和iDose算法(等级1～7)重建图像。对图像质量进行主客观评价,找出最佳的参数组合。对不同扫描参数和重建方法的剂量长度乘积(DLP)进行分析,计算有效剂量。结果在同一有效管电流下,iDose算法的噪声均低于FBP算法,且随着iDose等级的增加而下降;在不同的有效管电流下,噪声一般随着管电流的增加而下降。使用100mAs/层、iDose 5算法可以满足诊断要求,所获图像的噪声水平稍低于200mAs/层的FBP算法的图像。冠状位图像的可接受程度稍高于轴位图像。结论 iDose算法可满足颞骨CT图像质量的需求,又可降低患者的辐射剂量。

Application of iterative algorithm in temporal bone MSCT: An experimental study

Objective To investigate the influences of iterative algorithm (using iDose algorithm) on image quality and radiation dose in temporal bone MSCT. Methods A proper cadaveric head based on the noise level in routine clinical MSCT examination protocol with filtered back-projection (FBP) algorithm was selected as the research object. The cadaveric head was scanned with different effective tube current, and the images were reconstructed with FBP and iDose algorithms, respectively. The subjective and Objective evaluations were performed for all the images, and the optimal parameter combination was found. The values of CT volume dose index and dose length product (DLP) at various scan parameters and reconstruction algorithms were analyzed, and effective doses of subject were calculated as well. Results The noise level of iDose algorithm was lower than that of FBP under the same effective tube current, and it decreased with the increasing of iDose levels. The noise level decreased with the increasing of the effective tube current under different effective tube current. The combination of 100 mAs/silce and iDose 5 algorithm could met the requirements of clinical diagnosis, and the noise level of resultant images was lower than that obtained with the combination of 200 mAs/silce and FBP algorithm. The diagnostic acceptability of coronary images was slightly higher than that of axial images. Conclusion The iDose algorithm can not only meet the requirements of the temporal bone CT image quality, but also reduce the radiation dose of the patients.