光谱CT对头颈部金属植入物伪影的去除作用:一项基于人仿真模体的研究  

Reduction of Artifacts from Head and Neck Metal Implants by Spectral CT:A Study Based on Human Simulation Phantom

作　　者：崔永斌 黄勇 李若影 马长升 尹勇 CUI Yongbin;HUANG Yong;LI Ruoying(Department of Radiation Oncology,Shandong Cancer Hospital and Institute,Shandong First Medical University and Shandong Academy of Medical Sciences,Jinan,Shandong Province 250117,P.R.China)

机构地区：[1]山东省肿瘤防治研究院(山东省肿瘤医院)放射物理技术科,山东第一医科大学(山东省医学科学院),济南250117 [2]山东省肿瘤防治研究院(山东省肿瘤医院)影像科,山东第一医科大学(山东省医学科学院),济南250117

出　　处：《临床放射学杂志》2023年第12期1997-2006,共10页Journal of Clinical Radiology

基　　金：国家自然科学基金项目(编号:12275162)。

摘　　要：目的探究基于光谱CT虚拟单能图像(VMI)和矫正金属伪影去除技术(OMAR)对头颈部不同部位不同金属植入物伪影的去除作用。方法采用CIRS人仿真模体分别构建脊柱固定术后模型、金属假牙填充术模型、脑动脉瘤栓塞模型。并以金属不锈钢或者金属钛分别替代等效组织。利用Philips光谱CT分别扫描无金属植入物、金属不锈钢植入物、金属钛植入物条件下的三个部位模型。分别重组出60 keV、80 keV、100 keV、120 keV、140 keV、160 keV、180 keV以及120 kVp下的非OMAR图像和OMAR图像。测量金属植入物旁感兴趣区(ROI)内的CT值和SD值,并计算伪影指数(AI)和信噪比(SNR)。绘制出非OMAR和OMAR的各扫描条件下的AI值柱状图,以及SNR值变化表格,采用配对样本t检验比较差异。两个以上变量采用方差分析F检验进行比较。结果在颈部脊柱金属植入物模型中,金属钛造成的金属伪影最严重,特别是在VMI_(60)时AI值为160.103,SNR值为1.851;当能级为100 keV时,结合OMAR后AI值为4.689,SNR值为6.471。在金属假牙填充模型中,金属钛造成的金属伪影在VMI_(180)图像时AI值最小为12.831,OMAR处理后的图像AI值变化不显著(P=0.121),但是SNR值在VMI_(180)+OMAR时最大为2.262,明显优于VMI_(180)(P=0.002)。在脑动脉瘤栓塞术后模型中,AI值在VMI120及以上能级表现相当,经OMAR处理后,伪影值略降低,但SNR值明显增大。结论基于光谱CT的VMI图像和OMAR算法可以有效去除金属植入物产生的金属伪影,但并不是OMAR算法结合越高能级VMI图像去金属伪影能力就越好,其去除伪影的效力取决于金属植入物的材质以及部位。Objective To explore the effect of Virtual Monoenergetic Images(VMI)and Orthopedic Metal Artifact Reduction(OMAR)based on spectral CT on removing artifacts of different metal implants in different parts of the head and neck.Methods The models of spinal fixation,metal dental filling and cerebral aneurysm embolism were constructed with adult human phantom.Metal stainless steel and titanium are used to replace the equivalent structure.Philips spectral CT was used to scan the three site models under the conditions of metal-free implants,metallic stainless steel implants and metallic titanium implants.The non-OMAR images and OMAR images under 60 keV,80 keV,100 keV,120 keV,140 keV,160 keV,180 keV and 120 kVp were reconstructed respectively.The CT value and SD value in the ROI near the metal implant were measured,and the Artifacts Index(AI)and Signal-to-Noise Ratio(SNR)were calculated.Results In the spine metal implant phantom,the metal artifacts caused by metal titanium were the most serious,especially in VMI_(60),the AI value was 160.103,and the SNR value was 1.851;When the VMI energy level was 100 keV,the AI value after combining with OMAR was 4.689,and the SNR value was 6.471.In the metal denture filling model,the minimum AI value of metal artifacts caused by titanium was 12.831 in the VMI_(180)image,and the AI value of the image after OMAR treatment did not change significantly(P=0.121),but the maximum SNR value was 2.262 in the VMI_(180)+OMAR,which was significantly better than VMI 180(P=0.002).In the model after cerebral aneurysm embolization,the AI value was equivalent in the VMI 120 and above energy levels,and after OMAR treatment,The artifact value was slightly reduced,but the SNR value was significantly increased.Conclusion VMI image and OMAR algorithm based on spectral CT can effectively reduce metal artifacts caused by metal implants,but the higher VMI level combined with OMAR algorithm might not obtain the best effects,the effectiveness of removing artifacts depends on the material and site of metal implan

关 键 词：金属伪影 能谱成像 图像分析 金属植入物 伪影去除 

分 类 号：R816[医药卫生—放射医学]