【作者】 姬妮娜；

【导师】 高剑波；

【作者基本信息】 郑州大学 ， 影像医学与核医学， 2012， 硕士

【摘要】 背景和目的：胃癌是我国发病率最高的消化道恶性肿瘤,死亡率仅次于肺癌和肝癌,居第3位；而在农村其发病率及死亡率均居恶性肿瘤中的首位。同时胃也是胃肠间质瘤(Gastrointestinal stromal tumours, GISTs)及胃肠道淋巴瘤最常见的发病部位。MSCT增强扫描在胃肿瘤的诊断、鉴别诊断中的价值已得到大量研究的证实,但目前关于胃周动脉CT成像的研究相对较少。了解肿瘤的供血动脉及有无血管侵犯、有无血管变异等情况,对制定治疗方案及腹腔镜、开腹手术、介入治疗中避免血管损伤、减少术中出血、缩短手术时间具有重要意义。以往采用普通CT对胃周动脉成像的研究中,对管径较粗的胃左动脉及胃网膜右动脉的显示率几乎均达100%,但对管径较细的胃右动脉、胃后动脉、胃短动脉及胃网膜左动脉的显示率较低。宝石CT能谱成像技术(gemstone spectral imaging, GSI)在扫描过程中可以实现两种不同kVp射线的瞬时切换,从而获得两组吸收投影数据,并能重建出不同能量水平的单能量图像；其最佳对比噪声比(Optimal CNR)技术可以筛选出显示感兴趣区的最佳单能量图像,已有体外实验表明最佳单能量图像血管成像质量优于混合能量图像。本研究从以下三个方面探讨能谱CT在胃周动脉成像中的价值：一、能谱CT胃周动脉成像对胃周动脉解剖及变异的显示能力；二、能谱CT最佳对比噪声比(Optimal CNR)技术对胃周动脉成像质量的影响；三、能谱CT血管成像对胃肿瘤供血动脉的显示价值。材料与方法：1临床资料收集2011年5月至2011年12月临床怀疑胃部疾病而于本院行腹部能谱CT增强扫描的患者64例(排除有上腹部手术史者),男46例,女18例,年龄27-79岁,平均年龄61.2岁。手术或胃镜病理结果：胃癌45例、间质瘤9例(恶性伴肝转移1例、高度危险度3例、中等危险度1例、低危险度3例、极低危险度1例)、淋巴瘤6例、慢性胃炎4例。2扫描方法患者禁食8-12h,无禁忌症(如青光眼、前列腺增生等)者于扫描前10-15min肌肉注射盐酸山莨菪碱(654-2)20mg,并口服温开水800-1000ml。采用美国GE公司宝石CT(GE Discovery CT750HD)扫描,取仰卧位,扫描范围自肝顶至双肾下极。采用能谱扫描模式,使用高压注射器经肘前静脉注射碘海醇(欧乃派克,350mgI/ml),注射流率4ml/s,剂量1.5ml/kg；监测腹主动脉内CT值,达200HU时触发行动脉期扫描,于注射后60s行门脉期扫描。扫描参数：管电压为80kVp和140kVp(0.5ms瞬时切换),电流为自动毫安,层厚及层间距均为5mm,螺距0.984：1,转速0.8s/r,重建层厚及层间距均为0.625mm。3数据处理与分析数据传至GE ADW4.4工作站,由1位有经验的影像诊断医师利用GSI浏览器对动脉期薄层数据进行后处理。于腹腔干开口处取感兴趣区,取同层面之胃壁为对比,利用GSI Viewer中的Optimal CNR技术获得显示动脉的最佳单能量水平,保存该单能量水平的数据;取感兴趣区分别测量该单能量水平腹腔干开口处及同层面胃壁、腹壁皮下脂肪组织的CT值(每个部位测量3次,取平均值)；以相同方法测量140kVp混合能量图像中相应部位的CT值(保证两次测量所选图像的层面相同、感兴趣区的位置、大小、形状尽量保持一致)。利用公式CNR=|ROIo—ROId|/SDn,分别计算最佳单能量图像及混合能量图像中腹腔干与同层面胃壁的对比噪声比(contrast-to-noise ratio, CNR),作为胃周动脉成像图像质量的客观评价指标(其中ROIo为腹腔干开口层面的腹腔干CT值,ROId为同层面胃壁的CT值,SD。为同层面图像平均噪声,本研究采用同层面腹壁皮下脂肪组织CT值的标准差)。采用血管生长技术(add vessel)及薄层最大密度投影(thin slice maximum intensity projection, TSMIP)技术分别对最佳单能量图像及混合能量图像的胃左动脉(left gastric artery, LGA)、胃右动脉(right gastric artery, RGA)、胃网膜左动脉(left gastroepiploic artery, LGEA)、胃网膜右动脉(right gastroepiploic artery, RGEA)及胃短动脉(short gastric artery, SGA)、胃后动脉(posterior gastric artery, PGA)进行VR及TSMIP重建。由3位经验丰富的放射科医师分别独立对两组重建后的血管影像质量进行评价；同时观察并记录胃周动脉的显示率及其变异、胃肿瘤的部位、相应的供血动脉及供血动脉与肿瘤的关系。胃周动脉CTA (computed tomographic angiography)图像质量评价标准如下：4分：优,血管显示清晰,边缘光滑锐利；3分：良,血管主干显示尚清晰,末端显影较模糊；2分：中,血管显影较淡；1分：差,血管显影浅淡,仅能分辨走行。0分：未显影,血管结构不能分辨。对每例患者两组图像中的LGA、RGA、LGEA、 RGEA、PGA、SGA分别评分,取3位医师的平均评分作为最终评价结果。使用SPSS13.0软件进行统计分析,计算显示胃周动脉的平均最佳单能量水平,分别计算最佳单能量组及混合能量组VR图像中各支胃周动脉的显示率并采用确切概率法进行比较；对两组图像的腹腔干CT值、腹腔干对比噪声比、图像噪声、每支胃周动脉的主观评分及每例患者的总评分分别进行配对t检验(差值不服从正正态分布者,采用配对资料的符号秩和检验),以P<0.05为差异有统计学意义。结果：1.显示胃周动脉的最佳单能量水平为53±3keV,最佳单能量组及混合能量组VR图像对LGA及RGEA的显示率均为100.0%,PGA的显示率均为10.9%。LGEA的显示率分别为85.9%,78.1%；RGA显示率分别为71.9%,65.6%；SGA的显示率分别21.9%,15.6%,但采用确切概率法比较两组图像显示率差异无统计学意义(P值均大于0.05)。同时发现胃周血管变异35例,其中变异在2种或2种以上者14例。2.最佳单能量组的腹腔干CT值、腹腔干对比噪声比及图像噪声(654.68±120.17HU、36.96±11.76、16.80±4.73)与混合能量组图像(291.44±51.29HU、20.87±7.32、12.70±3.58)差异均具有统计学意义(P值均<0.001)；最佳单能量组分别较混合能量组增加约124.6%、77.1%、32.3%。3.最佳单能量组胃周动脉CTA中LGA、RGA、LGEA、RGEA、PGA、SGA的图像质量主观评分及总评分(3.73±0.44,2.39±1.55,2.35±0.72,3.48±0.63,0.20±0.70,0.60±1.09,12.75±2.81)与混合能量组(2.97±0.55,1.64±1.33,1.5±0.77,2.89±0.79,0.12±0.44,0.38±0.85,9.4±2.43)差异均有统计学意义(P值均<0.05),最佳单能量组较混合能量组分别增加约25.6%。45.7%。56.7%。20.4%。66.7%。57.9‰35.6%。4.60例肿瘤中19例局部供血动脉无明显异常改变,41例局部供血动脉增粗、迂曲、与肿瘤间脂肪间隙消失、被肿瘤包绕或穿入肿瘤；其中34例为胃癌患者,3例为淋巴瘤患者,4例为间质瘤患者。VR可以立体直观地显示供血动脉的起源、走行及增粗、迂曲等改变；TSMIP可以清晰地显示供血动脉与肿瘤的关系。结论：1.能谱CT胃周动脉成像可以清晰、直观地显示胃周动脉的解剖及变异；2.能谱CT最佳对比噪声比技术可以提高胃周动脉成像质量；3.能谱CT血管成像可以清晰、直观地显示胃肿瘤的供血动脉及其与肿瘤的关系。更多还原

【Abstract】 Background and ObjectiveGastric carcinoma is the most common malignant tumor of the digestive tract in China, and is the third leading cause of the cancer-related deaths, only inferior to lung cancer and hepatocellular carcinoma, and both the incidence and mortality of it is the first in malignant tumors in rural areas. And the stomach is the most common site of gastrointestinal stromal tumors(GISTs) and gastrointestinal lymphoma.The value of contrast-enhanced MSCT scan in diagnosis and differential diagnosis of gastric tumors has been proved by many researches, but the studies of gastric arterial CT angiography(CTA) are relatively fewer. To know the feeding arteries, vascular invasion and artery variation is essential for performing the best therapeutic regimen, avoiding vascular injury, reducing bleeding and shortening operation time.In the previous studies, the display rate of larger diameter arteries such as left gastric artery(LGA) and right gastroepiploic artery(RGEA) was nearly100%, but the display rate of smaller diameter arteries like right gastric artery(RGA), left gastroepiploic artery(LGEA), short gastric artery(SGA) and posterior gastric artery (PGA) was lower. Special imaging of GE Discovery CT750HD generates a set of single energy images. Recent study reveals that image quality of CTA of the optimal keV is better than that of the conventional polychromatic images.The aim of the study includes three aspects:First, to asess the value of special CT angiography on gastric arterial anatomy and artery variation. Second, to explore the effect of the optimal contrast-to-noise ratio(optimal CNR) technique on image quality of gastric arterial CTA in spectral imaging. Third, to evaluate the feeding arteries of gastric tumors.Materials and methods1Clinical datas64patients with clinically suspected gastric diseases from May2011to December2011were collected, including46males and18femals, excluding cases with histories of abdomen surgery. The range of age was between27and79, average at61.2. Pathological results:45cases were gastric cancer.9cases were gastric gastrointestinal stromal tumors(GISTs).6cases were gastric lymphoma, and4cases were chronic gastritis.2Scanning methodAll the subjects were asked to be fasting for8-12h, and the subjects without contraindications(such as benign prostate hypertrophy and glaucoma) were intramuscular injected654-220mg10-15min before the scanning, and were asked to drink water800～1000ml. All the patients took supine position, and underwent dual-phase enhanced gemstone spectral imaging examination, using a single tube, rapid dual kVp(80kVp and140kVp) switching technique on a high-definition Discovery CT750HD (GE Healthcare, Wisconsin, USA). The scanning range was from the top of liver to the inferior pole of kidney. The non-ionic contrast media Iohexol(Omnipaque350mgI/ml) at a dose of1.5ml/kg was injected with a high pressure syringe at a rate of4ml/s through median cubital vein. The arterial phase and portal venous phase scans were triggered when the aortic CT value was200HU and65s after the injection. The GSI scanning parameters:80kVp and140kVp switching per0.5ms, automatic tube current, the thickness and spacing:5mm, pitch:0.984:1, tube rotation time:0.8s. The reconstruction thickness was0.625mm, at an interval of0.625mm. 3Data processing and analysisData was sent to the GE ADW4.4workstation and data of the arterial phase was post-processed by an experienced radiologist using the GSI Viewer. The CNR plot of GSI viewer was used to select the optimal CNR for gastric artery. A circular region of interest(ROI) was placed within the base of the celiac trunk which we wanted to emphasize, and another ROI was placed within the gastric wall of the same slice which we wanted to contrast against. The GSI Viewer software package automatically calculated and displayed the CNR values for the101sets of monochromatic images, and from the CNR plot, the optimal monochromatic level for generating the best CNR between the gastric artery and the gastric wall could be selected. ROI was placed within the base of the celiac trunk, the gastric wall, and the subcutaneous fat tissue in the abdomen of the same slice to measure the CT value both in the optimal monochromatic images and in the140kVp polychromatic images. CNR was defined according to the formula: CNR=(ROIo-ROId)/SDn,(ROIo denoted CT value of the base of the celiac trunk, and ROId denoted the CT value of the gastric wall of the same slice, and SDn denoted the standard deviation of the subcutaneous fat tissue in the abdomen of the same slice). Measurements of all ROI were repeated for3times and the average value was chosen as the result.Both this monochromatic and the polychromatic image sets were post-processed with volume rendering(VR) and thin slice maximum intensity projection(TSMIP). Three experienced radiologists independently and subjectively graded overall noise and delineation of the left gastric artery(LGA), right gastric artery(RGA), left gastroepiploic artery(LGEA), right gastroepiploic artery(RGEA), short gastric artery (SGA) and posterior gastric artery(PGA) on CTA images with a4-point scale, and recorded the display rate of the arteries and the artery variation. Evaluation standard:4stood for excellent(arteries could be clearly shown and the margins were sharp);3stood for good(the arterial stem was clearly shown and the end was blurred);2stood for middle(the arteries were light);1stood for poor(only the distribution of the arteries were recognized);0arteries could not be recognized.Data was handled by SPSS13.0statistical package. The mean optimal keV for generating the best CNR was computed. Fisher’s exact probability was used to compare the display rate of gastric arteries on VR from the optimal monochromatic and the polychromatic images. Paired t-test or Wilcoxon matched-pairs signed ranks test was performed on CT value of the celiac trunk, contrast-to-noise ratio, image noise and subjective scores from the optimal monochromatic and the polychromatic images sets separately. P<0.05was defined as statistical significance.Results1. The monochromatic images at53±3keV was found to provide the best CNR for gastric arterial CTA. On monochromatic and the conventional140kVp VR images the display rate of LGA and RGEA were both100.0%; PGA, both10.9%; LGEA,85.9%and78.1%respectively; RGA,71.9%and65.6%; SGA21.9%and15.6%, but there was no significant difference(each P>0.05). At the same time, artery variation was found in35cases(in14cases, there was more than one kind of variation).2. The mean CT value, CNR of the celiac trunk and image noise at the optimal keV level(654.68±120.17HU,36.96±11.76,16.80±4.73) were higher, compared with those of the conventional polychromatic images(291.44±51.29HU,20.87±7.32,12.70±3.58)(each P<0.001). The mean CT value, CNR of the celiac trunk, image noise at the optimal keV level increased by124.6%,77.1%and32.3%respectively than those of the conventional polychromatic images.3. The subjective scores of the LGA, RGA, LGEA, RGEA, PGA, SGA and the total score of the CTA images at the optimal keV level(3.73±0.44,2.39±1.55,2.35±0.72,3.48±0.63,0.20±0.70,0.60±1.09,12.75±2.81) were all higher, compared with those of the conventional polychromatic images(2.97±0.55,1.64±1.33,1.5±0.77,2.89±0.79,0.12±0.44,0.38±0.85,9.4±2.43)(each P<0.05), increased by25.6%,45.7%,56.7%,20.4%,66.7%,57.9%and35.6%respectively.4. All the blood supply of60gastric tumors was displayed on CTA. In41cases, the feeding artery became dilatate; lost the adipose space, or the artery was wraped or penetrated into the tumors. VR can directly demonstrate anatomical structures and morphological changes of the arteries and TSMIP can clearly display the relationship between the arteries and the tumors. Conclusions1. Gastric arterial CTA of spectral CT can clearly and directly show the anatomy and vascular variation of the gastric arteries.2. The image quality of gastric arteries from the optimal monochromatic images was improved.3. Gastric arterial CTA of spectral CT can clearly and directly demonstrate the feeding arteries of the gastric tumors.更多还原