Quantitative and synthetic MR imaging

 描述（由申请人提供）：在接受典型MR检查时，患者可能需要躺在MR扫描仪中长达45分钟左右。在此期间，使用各种组织对比度对感兴趣的解剖结构进行重复成像。为了可视化任何图像特征，需要在该特征和背景组织之间存在显著的对比度，并且获取不同对比度的图像提高了能够可视化诊断相关特征的几率。目前提出的方法允许确定对比度的主要MR特性被快速地和定量地评估，使得随后可以计算而不是采集任何期望的对比度。根据在短短几分钟内采集的数据，本方法旨在生成所有主要物理MR参数的3D图：弛豫时间T1，T2和T2*，平衡磁化M0，翻转角和偏移频率Δf。 这里描述的方法的主要优点在于，每次一个地和/或通过线性方程来评估各种MR参数，而不是同时数值地求解涉及许多或所有参数的较大的非线性方程系统。可以使用不同的翻转角度a1和a2和/或不同的重复时间TR 1和TR 2采集少至两个MR数据集。虽然诸如TR的定时参数是精确已知的，但是另一方面，翻转角在空间上变化。从成像体积的中心开始，翻转角大约等于用户请求的值，翻转角以对象相关的方式逐渐减小，直到激发区域边缘处为零。为此，a1和a2必须被认为是要测量和评估的变量，而不是已知的成像参数。 基于MR参数图，可以计算而不是采集基本上任何MR对比度的3D图像，并且甚至可以发明新类型的对比度。这种几分钟的3D采集可能会取代整个检查。如果成功的话，这种方法可以用于筛查：人们可以以非常实惠的价格购买几分钟的扫描仪，没有人会死于发现太晚的肿瘤。此外，所提出的方法可能有助于加速MRI从定性到定量模式的转变，因为临床医生可能会逐渐喜欢查看MR定量图而不是定性对比。 目前的高风险高回报项目涉及为头部和关节（膝关节）成像定制拟议的3D定量映射方法。将在体模、志愿者和患者中验证该方法生成定量标测图以及常见造影剂类型合成图像的能力。对于头部和膝关节应用，该项目的主要目标是在相同的45分钟检查中从5名不同的健康志愿者获得诊断质量的3D图像，以证明该方法作为快速筛查方法的潜力。

 DESCRIPTION (provided by applicant): When receiving a typical MR exam, a patient may expect to lie in the MR scanner for as long as 45 min or so. During this time, the anatomy of interest gets repeatedly imaged, using a variety of tissue contrasts. To visualize any image feature there needs to be significant contrast between this feature and background tissues, and acquiring images of different contrast improves the odds of being able to visualize diagnostically- relevant features. The presently-proposed approach allows the main MR properties that determine contrast to be rapidly and quantitatively evaluated, so that any desired contrast might subsequently be computed rather than acquired. From data acquired in just a few minutes, the present method aims to generate 3D maps of all of the main physical MR parameters: The relaxation times T1, T2 and T2*, the equilibrium magnetization M0, the flip angle and the offset frequency Δf. A major strength of the approach described here is that the various MR parameters are evaluated one at a time and/or through linear equations, as opposed to numerically solving larger and non-linear systems of equations involving many or all parameters at once. As few as two MR data sets may be acquired, using different flips angles, a1 and a2, and/or different repetition times, TR1 and TR2. While timing parameters such as TR are precisely known, the flip angle, on the other hand, varies spatially. From the center of the imaged volume where it is approximately equal to the user-requested value, the flip angle progressively decreases all the way to zero at the edges of the excited region in ways that are object-dependent. For this reason, a1 and a2 must be thought of as variables to be measured and evaluated rather than known imaging parameters. Based on MR parameter maps, 3D images of essentially any MR contrast can be computed rather than acquired, and new types of contrast can even be invented. The proposed few-minute 3D acquisition might potentially replace a whole exam. If successful, the method could be used for screening: For a very affordable cost one might purchase a few-minute passage into a scanner, nobody should die of tumors discovered too late. Furthermore, the proposed approach may help accelerate the transition of MRI from a mostly qualitative to a quantitative modality, as clinicians in time may grow to prefer looking at MR quantitative maps rather than qualitative contrasts. The present high-risk-high-reward project involves tailoring the proposed 3D quantitative mapping approach for head and joint (knee) imaging. The ability of the method to generate quantitative maps as well as synthetic images of common contrast types will be validated in phantoms, volunteers and patients. For both the head and the knee application, a main goal of the project is to obtain 3D images of diagnostic quality from five different healthy volunteers in a same 45 minute exam, to demonstrate the potential of the approach as a fast screening method.