CT能谱成像在孤立性肺结节诊断中的应用

  本文关键词：CT能谱成像在孤立性肺结节诊断中的应用，，由笔耕文化传播整理发布。

        目的：通过对孤立性肺结节进行能谱成像(Gemstone Spectral Imaging, GSI)扫描,探讨能谱成像在肺结节鉴别诊断中的价值。材料与方法：1能谱CT对肺结节模型CT值测量的研究用直径不同的球体模拟肺结节(直径分别为20.0、10.0、5.0mm)；每种直径的结节分别由3种密度不同的物质制成(聚丙烯、尼龙、腈纶),相应CT值分别为-100HU、60HU、100HU。结节固定在塑料肺基底模型(-900HU)内,周围加基液或基液加动物骨骼以改变模型外周环境。Discovery CT750HD能谱CT对肺结节模型进行GSI扫描和常规扫描(混合能量)。GSI扫描参数：140kVp/80kVp瞬时切换；630mA;层厚64×0.625mm；螺距0.516：1；SFOV50cm;DFOV38cm。常规扫描管电压120kVp,管电流(mA)根据GS工扫描患者的CT剂量指数确定,其他扫描参数与GSI扫描一致。对两种不同外周环境中的模型分别用三种不同的转速(0.5s/r,0.8s/r,1.0s/r),即扫描时间分别为(0.5s,0.8s,1.0s)进行扫描。1.1三位经验丰富的影像学医师独立测量不同扫描条件下直径为20mm结节的CT值及SD值。混合能量和单能条件下所得结节的CT值及相应的图像信噪比SNR进行比较。1.2应用能谱图像分析软件GSI-Viewer对所有结节的单能量图像进行数据测量及分析,得出相应的最佳CNR曲线。三位经验丰富的影像学医师独立测量并记录每个结节的最高CNR的keV单能量,测量3次,取其平均值。最佳CNR的KeV条件下测量直径为20mm的结节模型CT值与73KeV及实际值比较分析。1.3在最佳CNR KeV单能图像上对所有肺结节模型单能量图像进行数据测量及分析,得出相应的能谱曲线。反映CT值衰减的CT能谱曲线的差异可以用曲线斜率来定量评估,分析其与结节的密度、直径、扫描时间、外周环境的线性相关性。2.能谱CT对肺结节患者扫描Discovery CT750HD能谱CT对42例经病理证实或随访的孤立性肺结节患者进行GSI扫描。扫描参数：140kVp/80kVp瞬时切换；630mA；层厚64×0.625mm；螺距0.516:1;SFOV50cm；DFOV38cm；转速0.8s/r。其中鳞癌17例,腺癌15例,肺炎10例。应用能谱图像分析软件GSI-Viewer对单能量图像进行分析,得出能谱曲线,分析不同病理类型的肺结节能谱曲线的差异性。结果：1能谱CT对肺结节模型扫描1.1(1)分别在基液与基液+骨骼条件下对肺结节模型进行GSI扫描：所得CT值均无明显统计学差异(P>0.05)。所有结节转速为0.8s/r及60Hu结节转速为1.0s/r时SNR无明显统计学差异(P>0.05)；其他条件下SNR间差别有统计学意义(P<0.05)。(2)分别在基液与基液+骨骼条件下对肺结节模型进行常规扫描：所得CT值间差别均有统计学意义(P<0.05)。所有SNR间差别亦有统计学意义(P<0.05)。1.2(1)所有结节模型所测得最佳CNR符合正态分布,其95%可信区间为(64.75-65.27),则在均值65keV时,能获得最高的CNR。(2)65keV.73keV时测得直径为20mm结节的CT值与实际CT值的差异及65keV与73keV之间差异均有统计学意义(P<0.05)。GSI扫描73keV条件下-100HU、60HU、100HU结节模型CT值分别为-133.95±0.60、23.47±1.04、89.33±1.38；65keV条件下CT值分别为-119.51±2.43、40.39±1.43、102.63±1.88。即73keV比65keV时测得结节CT值的平均值均更接近于实际的CT值。1.3肺结节的密度、直径、扫描时间、外周环境作为自变量,能谱曲线CT值衰减率作为因变量建立一个多元回归线性模型。结果显示：“能谱曲线CT值的衰减率”与“结节密度”、“结节直径”呈正相关；“结节密度”与“能谱曲线CT值的衰减率”均比“结节直径”与“能谱曲线CT值的衰减率”具有更强的线性相关性。2能谱CT对孤立性肺结节患者扫描对鳞癌、腺癌、肺炎的能谱曲线CT值衰减率的比较方差分析结果示：F=167.727,P=0.000<0.005,处理因素的3组水平中至少有一组的总体平均值不同于其他各组。鳞癌、腺癌、肺炎两两比较,P均<0.005,可以认为鳞癌组与腺癌组、鳞癌组与肺炎组、腺癌组与肺炎组的能谱曲线CT值的衰减率不等,差异有明显统计学意义。结论：GSI扫描可有效减除射束硬化伪影对CT值的影响得到较常规扫描更准确的肺结节CT值,肺结节模型在65keV的单能水平下得到最高CNR。CNR。尽管73keV条件下GSI扫描测得的CT值与实际CT值间有一定差异,但较65keV条件下CT值更准确。肺结节能谱曲线CT值的衰减率与结节密度、结节直径呈正相关,结节密度相关性更大。即不同密度、不同直径的结节,其能谱曲线CT值的衰减率有差异一定,不同密度结节之间差异大。不同病理类型的肺结节其能谱曲线CT值的衰减率有一定差异。

    ObjectiveTo evaluate the diagnosis of gemstone spectral imaging (GSI) of solitary pulmonary nodules with Discovery CT750HD.Methods1. A pulmonary nodule phantom imaging with Discovery CT750HD:The spherical pulmonary nodules made with three different kinds of materials (Polypropylene, Monomer casting nylon and ACRYLIC). Each material was applied to represent different density (-100HU,60HU and100HU). The diameters of nodes were20.0mm,10.0mm,5.0mm. All the nodules were set in a plastic lung phantom (-900HU). Phantom was fixed in a container filled with fluid, muscle and bone to simulate thorax. The phantom scan was performed under both single energy (SE) and mixed energy (ME) CT techniques. Scans were performed with Discovery CT750HD scanner using GSI (Gemstone Spectral Imaging) technique (GSIKV,630mA,64×0.625mm slices,0.516:1pitch,50cm scan field of view,38cm display field of view) with there different ratation time (0.5s,0.8s,1.0s). In addition, images were generated using the high resolution120kVp technique (convert mA according to GSI imaging CTDI, other parameters were the same with SECT). Scans were performed with and without the added bone.1.1Quantitative CT numbers and SD numbers measurement of nodules (diameters of nodes were20.0mm) at different scan conditions were done by three experienced radiologists on the workstation. CT numbers SNR of each nodule under different SECT and MECT were compared.1.2Using the spectral analysis software to acquire the best contrast-to-noise ratio (CNR) curves of all nodules. The homologous keV of of the best CNR measurement of nodules were done by three experienced radiologists on the workstation, repeated three times and took the averages. Comparing the CT numbers of the homologous keV of best CNR with the CT numbers of73keV, the diameter of nodes was20.0mm.1.3Using the spectral analysis software to acquire the spectrum curves of all nodules. The curves slope could evaluate the difference of spectrum curves. Analysising the Linear correlation between the curves slope and nodules density, nodules diameter, ratation time and added bone.2. Pulmonary nodule patients imaging with Discovery CT750HD:GSI of42pulmonary nodule patients (squamous carcinoma17, adenocarcinoma15, pneumonia10) confirmed by pathology or foliowup data. Scans were performed with Discovery CT750HD (GSIKV,630mA,64×0.625mm slices,0.516:1pitch,50cm scan field of view,38cm display field of view,0.8s ratation time). Analysising the differences of spectrum curves between different pulmonary nodules pathology.Results1. A pulmonary nodule phantom imaging with Discovery CT750HD:(1) The phantom scanned under SE in Nacl+Model or Nacl+Model+Bone with Discovery CT750HD:All the CT numbers didn’t show significant difference (P>0.05). In conditions of (0.8s, ratation time) and (1.0s, ratation time;60HU, density), the SNR didn’t show significant difference (P>0.05), the other conditions the SNR showed significant difference (P<0.05).(2) The phantom scanned under ME in Nacl+Model or Nacl+Model+Bone with Discovery CT750HD:All the CT numbers showed significant difference (P <0.05). All the SNR also showed significant difference (P<0.05).1.2(1) The CNR of pulmonary nodules met normal distribution. The confidence interval was (64.75-65.27).65keV image of pulmonary nodule and lung had the best CNR.(2) The CT numbers of65keV,73keV and standard CT numbers of nodules (diameters of nodes were20.0mm) show significant difference (P<0.05). The CT numbers of65keV and73keV also show significant difference (P<0.05). The CT numbers of-100HU,60HU,100HU were-133.95±0.60,23.47±1.04,89.33±1.38under73keV, and were-119.51±2.43,40.39±1.43,102.63±1.88under65keV. The CT numbers of73keV were more accurate.1.3Taking the nodules density, nodules diameter, ratation time, added bone as independent variable and taking the spectrum curves slope as dependent variable to acquire multiple regression linear model."The spectrum curves slope" was positively related with "nodules density","nodules diameter". And "nodules density" had more strong linear correlation than "nodules diameter".2. Pulmonary nodule patients imaging with Discovery CT750HD:Analysis of variance showed:F=167.727, P=0.000<0.005. Comparing any two kinds of pulmonary nodules pathology (squamous carcinoma, adenocarcinoma, pneumonia) showed P<0.005. So the spectrum curves slope had a marked difference between different pulmonary nodules pathology.ConclusionThe Discovery CT750HD results indicated that SECT can alleviate beam hardening effect, acquire more accurate CT numbers than MECT, and get the best CNR at65keV. The CT numbers of73keV had some differences with standard CT numbers, but were more accurate than65keV. The spectrum curves slope of pulmonary nodules was positively related with nodules density, nodules diameter, nodules density was more obviously. The spectrum curves slope of pulmonary nodules also had differences between different pulmonary nodules pathology.

          CT能谱成像在孤立性肺结节诊断中的应用

中文摘要6-9Abstract9-11符号说明12-13前言13-15第一部分15-28    材料与方法15-17    结果17-21    讨论21-28第二部分28-32    材料与方法28    结果28-29    讨论29-32结论32-33附图33-38参考文献38-42综述42-58    综述参考文献50-58致谢58-59攻读学位期间发表的学术论文59-60学位论文评阅及答辩情况表60

  本文关键词：CT能谱成像在孤立性肺结节诊断中的应用，由笔耕文化传播整理发布。