临床外科杂志 ›› 2024, Vol. 32 ›› Issue (3): 301-305.doi: 10.3969/j.issn.1005-6483.2024.03.022

• 论著 • 上一篇    下一篇

基于模型的迭代重建、金属伪影去除和噪声优化的虚拟单能成像在去除腰椎内固定金属伪影效果的正交设计研究

成庆 夏可周 彭宙峰 程亮 郭卫春   

  1. 430060 武汉大学人民医院脊柱Ⅰ科(成庆、夏可周、郭卫春),放射科(彭宙峰、程亮)
  • 收稿日期:2023-11-20 出版日期:2024-03-20 发布日期:2024-03-20
  • 通讯作者: 郭卫春,Email:guoweichun@aliyun.com

The Effect of model-based iterative reconstruction,metal artifact reduction,noise-optimized virtual monochromatic imaging in removing lumbar internal fixation metal artifacts:an orthogonal design study

CHENG Qing,XIA Kezhou,PENG Zhoufeng,CHENG Liang,GUO Weichun   

  1. Department of Spine Ⅰ,Renmin Hospital of Wuhan University,Wuhan 430060,China
  • Received:2023-11-20 Online:2024-03-20 Published:2024-03-20

摘要: 目的 比较基于模型的迭代重建(MBIR)算法、金属伪影去除(MAR)算法和噪声优化的虚拟单能成像(VMI+)在去除腰椎内固定金属伪影的效果差异,并提出最佳的联合方案。方法 回顾性收集我院腰椎内固定术后病人的腰椎CT原始数据,共10例,通过正交实验设计,设计3因素2水平的正交表L8(27),将原始数据按照正交设计表的方案行重建后处理,分别测量螺钉邻近侧腰大肌处高密度金属伪影区域、螺钉长轴腹侧低密度金属伪影区域以及背部皮下脂肪对照区域的CT值和SD值,并计算AI,随后进行直观分析和方差分析。结果 在对高密度金属伪影的影响中,因素的影响主次为:C(VMI+)>B(MAR)>A(MBIR);在对低密度金属伪影的影响中,因素的影响主次为:B(MAR)> C(VMI+)> A(MBIR)。各因素不同水平对高密度和低密度金属伪影的影响均为A2>A1、B2>B1、C2>C1。A(MBIR)、B(MAR)、C(VMI+)是去除腰椎内固定高密度和低密度金属伪影的显著影响因素(P<0.001)。在对高密度金属伪影的影响中,A与B、A与C、B与C的交互作用显著(P<0.001);在对低密度金属伪影的影响中,A与B、B与C的交互作用显著(P<0.05)。最佳的联合方案为A2B2C2。结论 MBIR、MAR、VMI+是去除腰椎内固定金属伪影的显著因素。MBIR、MAR、VMI+在去除腰椎内固定金属伪影的效果上存在差异:对高密度金属伪影,VMI+>MAR>MBIR;对低密度金属伪影,MAR> VMI+> MBIR。MBIR、MAR、VMI+在联合应用于去除腰椎内固定金属伪影中,存在交互作用,其中以MAR和VMI+的交互作用最为显著。本研究的最佳的联合方案为MBIR 4级、130KeV VMI+、MAR。

关键词: MBIR;MAR;VMI+;腰椎内固定;金属伪影;正交设计

Abstract: Objective To compare the efficacy differences among model-based iterative reconstruction (MBIR),metal artifact reduction(MAR),and noise-optimized virtual monochromatic imaging(VMI+) in the removal of metal artifacts from lumbar spine internal fixation and to propose the best combined scheme.Methods Retrospective collection of lumbar CT raw data of patients after lumbar internal fixation surgery in our hospital,a total of 10 cases.Through orthogonal experimental design,a 3-factor 2-level orthogonal table L8 (27) was designed.The raw data was reconstructed and post-processed according to the scheme of the orthogonal design table,and the CT values and SD values of the high-density metal artifact area near the screw,the low-density metal artifact area on the ventral side of the screw long axis,and the subcutaneous fat control area on the back were measured separately.The Artifact Index (AI) was calculated,followed by intuitive analysis and analysis of variance (ANOVA).Results In the impact on high-density metal artifacts,the order of influence of factors was:C (VMI+) > B (MAR) > A (MBIR);for low-density metal artifacts,the order was:B (MAR) > C (VMI+) > A (MBIR).The effects of different levels of each factor on both high-density and low-density metal artifacts were A2 > A1,B2 > B1,C2 > C1.A (MBIR),B (MAR),and C (VMI+) were significant influencing factors in the removal of high-density and low-density metal artifacts from lumbar spine internal fixation (P<0.001).The interactions between A and B,A and C,and B and C were significant in the impact on high-density metal artifacts (P<0.001);for low-density metal artifacts,the interactions between A and B,and B and C were significant (P<0.05).The optimal combined protocol was identified as A2B2C2.Conclusion Conclusion MBIR,MAR,and VMI+ are significant factors in the removal of metal artifacts from lumbar spine internal fixation.There are differences in the effects of MBIR,MAR,and VMI+ on the removal of metal artifacts:for high-density artifacts,VMI+ > MAR > MBIR;for low-density artifacts,MAR > VMI+ > MBIR.There is an interaction when MBIR,MAR,and VMI+ are used in combination to remove metal artifacts from lumbar spine internal fixation,with the interaction between MAR and VMI+ being the most significant.The best combined scheme of this study is MBIR level 4,130KeV VMI+,MAR.

Key words: model-based iterative reconstruction; metal artifact reduction; noise-optimized virtual monoenergetic imaging; lumbar internal fixation; metal artifacts; orthogonal design

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!