【作者】 王海燕；

【导师】 赵斌；

【作者基本信息】 山东大学 ， 影像医学与核医学（专业学位）， 2014， 博士

【摘要】 第一部分不同级别SAFIRE重建对常规剂量胸部CT图像质量及噪声的影响研究背景：CT是胸部疾病最常用的检查方法,随着CT在临床上应用持续增加,CT检查导致的辐射剂量也不断增加。CT检查应该遵循剂量最低化原则,即在保持足以满足诊断需要的较好的图像质量的同时,最大程度地降低辐射剂量,这是目前CT技术研究的方向和焦点之一。在降低CT检查的辐射剂量方面,相关研究人员开发了很多新技术,如在CT扫描环节降低管电压及管电流,根据受检者体重及层厚采用个性化扫描参数,通过自动管电压及管电流调节技术降低扫描剂量,大螺距扫描技术等。在CT图像重建环节,应用传统滤波反投影法(FBP, Filtered Back Projection)进行图像重建时,CT辐射剂量的降低将会带来图像噪声增加,导致图像质量降低,这种制衡关系限制了CT辐射剂量的进一步降低,迭代重建的使用可以克服滤波反投影算法辐射剂量与图像质量之间的制衡关系,可以更大程度降低扫描的辐射剂量,同时维持较优质的图像质量。近年来各个CT生产厂家推出了多种不同的迭代重建技术应用于临床。基于原始数据迭代重建(SAFIRE, Sinogram-aff irmed iterative reconstruction)是最新的迭代重建技术之一,在构建模型上是基于原始数据的迭代重建算法,能够显著降低图像噪声、减少伪影,有效提高图像质量,从而有可能进一步减少辐射剂量。研究目的：行常规剂量胸部CT扫描后,图像数据应用传统滤波反投影重建算法(FBP)及不同级别SAFIRE迭代重建算法分别进行图像重建,对图像噪声,信噪比及主观图像质量进行对比,探讨获得最佳图像质量的迭代重建级别。研究方法：应用双源FLASH CT对30例成年人受检者行胸部CT扫描。具体扫描条件如下：所有受检者均应用西门子双源Flash CT (Somatom Definition Flash, Siemens Healthcare, Forchheim, Germany)采用双球管单能模式进行胸部CT平扫,管电流：120kV,开启CARE Dose4D,准直64x2x0.6mm,螺距1.2,球管旋转时间0.5s,层厚及间距均为5mm。扫描范围为吸气后单次屏气从胸廓入口层面扫描至肺底层面。对扫描后的原始数据分别采用传统应用的滤波反投影重建算法(FBP)和1-5级别基于原始数据的迭代重建(S AFIRE)技术进行重建。得到6组不同重建技术的图像资料,测量同一层面不同重建技术重建图像的噪声、CT值,计算信噪比(SNR, signal noise ratio)并对图像噪声及信噪比进行统计学分析。另外对各组肺窗及纵隔窗图像进行主观图像质量评分,由两名医学影像诊断医师采用双盲法以5分法对图像质量进行评分并进行统计学分析。结果：SAFIRE1-5迭代重建图像的噪声均低于FBP重建图像,图像SNR均高于FBP重建图像。随着SAFIRE迭代重建权重增加,图像噪声逐渐降低,图像SNR逐渐增高,呈线性相关关系。SAFIRE重建图像中,SAFIRE1噪声最大、信噪比最小,SAFIRE5噪声最小、信噪比最大,SAFIRE2-4噪声及SNR数值居中。CT值对比各组数据之间均未见统计学差异。SAFIRE3、4重建图像平均主观评分最高。结论：在胸部CT扫描中,与常规重建算法滤波反投影重建(FBP)比较,采用基于原始数据迭代重建技术(SAFIRE)能显著提高图像信噪比,降低图像噪声,随着迭代级别增高,噪声及信噪比改善程度逐渐增加。权重3、4时图像主观评分最高。第二部分低剂量胸部CT图像质量：SAFIRE与FBP对照研究研究背景：CT在临床上的广泛应用随之带来辐射剂量的增加,目前CT低剂量技术研究成为CT研究的热点之一。应用常规滤波反投影重建算法(FBP), CT扫描辐射剂量的降低就会带来CT图像噪声的升高及CT图像质量的降低,这种制衡关系限制了CT剂量的进一步降低。在CT图像重建环节,迭代重建可以显著降低CT图像的噪声,提高图像质量,可以解决传统滤波反投影算法中剂量降低与图像质量之间的制衡关系。迭代重建早期出现时因重建时间长和运算构建复杂等缺点而难以在临床中广泛应用。近年来随着技术发展迭代重建时间缩短到在临床可接受范围。相比滤波反投影算法,迭代重建算法可以选择性去除图像噪声,获得较好的图像质量,从而进一步降低CT扫描的辐射剂量。SAFIRE重建是最新推出的迭代重建算法之一,是基于原始数据的迭代重建算法。研究目的：常规剂量胸部CT扫描行FBP重建作为对照组,低剂量胸部CT分别行SAFIRE及FBP重建,与对照组进行对照,对比图像噪声,信噪比及主观图像质量及病变的显著性显示。探讨SAFIRE重建降低图像噪声、提高图像质量的潜力,寻求较优的低剂量扫描方案。研究方法：87例受检者(46名男性,41名女性,平均年龄为54.54±16.12,平均体重指数24.58±4.07)初次检查行常规剂量胸部CT扫描(平均DLP183.37±44.13mgy. cm),随访时胸部CT行降低50%剂量的低剂量胸部扫描(平均DLP91.08±91.08mgy. cm)常规剂量扫描进行FBP重建,低剂量扫描分别行基于原始数据的SAFIRE迭代重建与FBP重建。不同重建图像在相同图像层面测量图像噪声与信噪比(SNR),共测量四个感兴趣区：气管分叉水平升主动脉、降主动脉,纵隔窗气管分叉上方气管区域,肺窗气管分叉上方气管区域,3组数据的噪声及信噪比应用配对t检验进行统计学分析。采用5分级法进行主观图像质量评估,主观图像质量和病变显著性的比较应用wilcoxon符号等级检验进行对比。结果：随访低剂量扫描的实际辐射剂量为初始剂量的大约50%(49.26%±2.62)。与常规剂量FBP重建图像对照,低剂量SAFIRE重建纵隔窗图像噪声(肺动脉分叉水平升主动脉：P=0.38,降主动脉：P=0.70,纵隔窗气管：P=0.37)与图像信噪比(肺动脉分叉水平升主动脉：P=0.14,降主动脉：P=0.72,纵隔窗气管P=0.06)无显著统计学差异,肺窗图像噪声显著降低(P<0.001),信噪比显著增高(P<0.001)；低剂量FBP重建图像图像噪声显著增高(P<0.001),图像信噪比显著降低(P<0.001)。与常规剂量FBP重建图像对照,低剂量SAFIRE重建纵隔窗(p=0.317)与肺窗(p=0.614)的主观图像质量与病变显著性均未见显著性差异；低剂量FBP主观图像质量在纵隔窗(p<0.001)与肺窗(p<0.001)均显著性降低,部分病变显著性降低(磨玻璃影,p<0.0001；边界不清小结节,p<0.0001；肺囊肿,p<0.0001；肺气肿,p=0.003)。结论：与常规剂量胸部CT扫描FBP重建图像比较,减少50%的剂量低剂量胸部CT应用SAFIRE重建图像可以提供类似或更好的图像质量。低剂量FBP重建图像质量显著性减低。更多还原

【Abstract】 Part1Application of different level iterative reconstruction of routine dose chest CT on image quality and noiseBackground:CT is the most commonly used inspection method of chest disorders. However, along with increasing clinical application of chest CT examination, the radiation exposure that associated with CT has been raised. Accordingly, CT examination should follow the ALARA (As Low As Reasonably As Achievable) principle. The main challenge of CT is to keep radiation dose to a minimum, while still obtaining better diagnostic image quality that enough to meet the needs of diagnosis. Reducing the radiation dose to the greatest extent is the focus of CT technology research. In order to reduce CT radiation dose, the relevant researchers have developed a lot of new image acquisitions, such as selecting weight-adapted protocols, using automatic tube voltage and tube current modulation systems, and large scan pitch technology, etc.Using traditional FBP reconstruction algorithm, there was a trade-off between the downside of radiation dose and image noise. The reduction of CT radiation dose will lead to reduction of image noise and image quality, this kind of trade-off relationship limits the further reduction of radiation dose of CT. This trade-off limits the minimum radiation dose required for a specific diagnostic application. Recently, the iterative reconstruction techniques were another option to save dose. The use of the iterative reconstruction can overcome the limit of filter back projection algorithm between radiation dose and image quality. Some studies about iterative reconstruction were introduced to improve image quality, enhance image resolution and lower image noise. Many CT manufacturers have launched different iterative reconstruction technologies that were applied in clinical in the past few years. Sinogram-affirmed iterative reconstruction (SAFIRE) is one of recently introduced iterative reconstruction algorithm, which is aimed at overcoming the constraints of conventional FBP by decoupling the relationship between spatial resolution and image noise. SAFIRE estimates the local noise content in each image pixel by analysis the raw data contributing to this pixel, and removed it from the current image data set. SAFIRE can significantly reduce the image noise and reduce artifacts, effectively improve the quality of the image, which could further reduce the radiation dose.Purpose:To evaluate the application of different level of SAFIRE reconstruction technique of routine dose chest CT examination on subjective image quality and image noise, SNR (signal noise ratio) and compared image quality with conventional FBP reconstruction image, to explore the iterative reconstruction level of best image quality and the lowest noise.Material and Methods:The routine dose chest CT examinations of30patients were performed on the dual source flash128-slice MDCT system (SOMATOM Definition Flash, Siemens Healthcare, Germany). All were unenhanced CT examinations. The scopes of CT examinations were from the thoracic inlet to the diaphragm during a breath hold after inspiration. Using two tubes with single energy mode, with the following parameters as collimation:64×2×0.6mm with z-flying focal spot; CARE Dose4D; pitch:1.2; rotation time:0.5s; section thickness and inter slice spacing:5mm. The initial lung and mediastinal images reconstructed with FBP algorithm and SAFIRE1-5algorithm respectively. In each data set, the objective assessment of noise was obtained by measuring the standard deviation of the measured Hounsfield units (HU) within circular regions of interest (ROIs) on axial images by a radiologist with5years experience in CT. The signal attenuation was measured on the same time, and the signal-to-noise ratio (SNR) was calculated. Subjective image quality was rated in5point Likert scale by two radiologists with12and16years experience in CT, respectively, who blinded to reconstruction technique. On each series of mediastinal and lung images, subjective noise was rated using a5-point Likert scale according to the severity of image noise, quality of contour delineation, and general image impression.(1=excellent;2=good;3=moderate;4= fair;5=poor). The final scores of subjective image quality were decided by consultation by two radiologists.Result:Noises of SAFIRE1-5were all lower than those of FBP, and SNR values were all higher than those of FBP. As SAFIRE iterative reconstruction strength increased, gradually reduce the image noise, improving the image SNR. The comparison of CT values has no significant statistical difference. SAFIRE2-4reconstruction image have the highest subjective score.Conclusion:On chest CT examination, compared with conventional filter back projection (FBP) reconstruction algorithm, the application of iterative reconstruction technique (SAFIRE1-5) can significantly improve image quality and reduce the image noise, improve signal to noise ratio. Using SAFIRE2-4can get the best subjective image quality. Part2Comparison of SAFIRE versus FBP reconstruction of low-dose chest CT on image noise and image qualityBackground:Computed tomography (CT) is widely used to bring the increase of radiation dose. Low dose technique became the hot topics in the study of CT technology. CT radiation dose reduction would bring increasing of image noise used conventional FBP reconstruction. This trade-off limits the further reduction of CT dose. Iterative reconstruction can effectively reduce the noise in the reconstruction of the images, can be solved the trade-off limits in traditional filter back projection of FBP algorithm. In recent years, along with the technology development of computer, the time of iterative reconstruction was shortened to clinical acceptable range. Compared to FBP algorithm, the iterative reconstruction algorithm can significantly reduce the scanning dose of selectively remove the noise of the image at the same time, in order to gain a better image quality. SAFIRE reconstruction is one of the latest iterative reconstruction algorithms that based on the original data.Objective:Routine dose chest computed tomography (CT) reconstructed with filtered back projection (FBP). Low dose chest computed tomography (CT) reconstructed with FBP and with raw data-based iterative reconstruction. To compare image noise, image SNR, subjective image quality and lesion conspicuity, and to evaluate the ability of SAFIRE reconstruction technology that reduce the scanning radiation dose and find the best scan protocol.Methods:87consecutive patients (46male, mean age54.54±16.12, mean BMI24.58±4.07) referred for initial chest CT with full-dose examinations (mean DLP183.37±44.13mGy.cm) and follow-up chest CT with half-dose examinations (mean DLP91.08±23.81mGy.cm). The full-dose protocol reconstructed with FBP, half-dose protocol reconstructed with FBP and Sinogram-affirmed iterative reconstruction (SAFIRE) respectively. Noise and signal-to-noise (SNR) were compared using a paired Student t-test, subjective image quality and lesion conspicuity was compared using wilcoxon signed ranks test. Results:Actual radiation dose of follow-up CT was about50%(49.26%±2.62) of routine dose protocol. Compared to full-dose images with FBP, there was no significantly difference in half-dose images with SAFIRE in the objective noise (Ascending aorta:P=0.38, Descending aorta:P=0.70, Trachea on mediastinal images: P=0.37) and SNR (Ascending aorta:P=0.14, Descending aorta:P=0.72, Trachea on mediastinal images:P=0.06) on mediastinal images; Noise was significantly lower (P<0.001) and SNR was significantly higher (P<0.001) in half-dose images with SAFIRE on lung images; Noise was significantly higher (P<0.001) and SNR was significantly lower (P<0.001) in half-dose images with FBP. Subjective image noise and lesion conspicuity were both similar on mediastinal images (p=0.317) and lung images (p=0.614) of half-dose SAFIRE images versus full-dose FBP images. Subjective image quality were both significantly lower on mediastinal images (p<0.001) and lung images (p<0.001) of half-dose FBP images versus full-dose FBP images. The conspicuity of some lesion was significantly lower (ground-glass opacitie, p<0.0001; ill-defined micronodule, p<0.0001; lung cyst, p<0.0001; emphysematous lesion, p=0.003) on half-dose FBP versus full-dose FBP images.Conclusion:Compared to full-dose CT images reconstruction with the conventionally FBP algorithm, SAFIRE with three iterations could provide similar or better image quality on at50%less dose. Half-dose FBP images provide poor image quality at50%less dose.更多还原