FAST DYNAMIC 3D MRI USING ADAPTIVE SPATIAL ENCODING

描述（改编自申请人的摘要）：本发明的主要目标是： 该项目是一种快速动态3D MRI方法的发展，使用自适应 可采集高分辨率MRI数据的空间编码（256 x256 x256） 尽可能接近实时操作（每1-2秒1个数据集）， 对标准MRI扫描仪进行最小的硬件修改。 这个目标在于 远远超出了当前MRI方法的可能性， 采用傅立叶编码和专用梯度硬件的技术。 介入性磁共振成像的一些应用，是我院的一个重点， 具有动态3D MRI的特定要求，可以在 没有专门梯度线圈的“开放式”MR扫描仪。 最重要的 这些应用是MRI监测热传导的时间过程， 在治疗过程中由于组织 异质性和附近的血管。 另一个重要的应用是近 用于微创手术的探头和导管的实时3D跟踪 治疗 具体而言，我们建议开发、实施、测试和 优化在两个方向上自适应编码的动态3D MRI方法 使用高翻转角2D空间选择性RF激励来实现 最小的一组接近最优的编码计算从多维 3D图像估计的奇异值分解（MSVD）（由 最近采集的数据），结合频率 在第三方向上编码。 主要目标的实现 该项目之所以能够成功，是因为三个重要的技术 预付款。 首先，关于空间编码， 申请人最近报告说，他们开发了MSVD，一种强大的数值计算方法， 这是一个数学工具，可以确定接近最佳的3D空间编码。 第二，一个简单的快速数值程序已经开发，在他们的 实验室用于计算RF脉冲波形， 非傅立叶编码使用高翻转角（90=A1）进行高SNR扫描。 第三，在他们的设施，他们有能力为近 实时自适应2D MRI，使用商业MR系统， 修改附加的工作站。

DESCRIPTION (Adapted from Applicant's Abstract): The main goal of this project is the development of a fast dynamic 3D MRI method using adaptive spatial encoding that can acquire a high resolution MRI dates (256x256x256) operating in near real-time as possible (1 dataset per 1-2 seconds) with minimal hardware modifications to a standard MRI scanner. This goal lies well outside the possibilities of current MRI methods like echo planar techniques that employ Fourier encoding and specialized gradient hardware. A number of applications of interventional MRI, a focus in our hospital, have the specific requirement for dynamic 3D MRI that can operate on an "open" MR scanner with no specialized gradient coils. The most important of these applications is the MRI monitoring of the timecourse of thermal therapies during which non-uniform heating of tissue occurs due to tissue heterogeneity and nearby vessels. Another important application is the near real-time 3D tracking of probes and catheters used for minimally invasive therapies. Specifically, we propose to develop, implement, test and optimize a dynamic 3D MRI method that encodes adaptively in two directions using high flip angle 2D spatially selective RF excitations to implement a minimal set of near-optimal encodes computed from the multidimensional Singular Value Decomposition (MSVD) of a 3D image estimate (formed from recently acquired data) computed per acquisition, combined with frequency encoding in the third direction. The accomplishment of the main objective of this project is only possible due to three significant technological advances. First, and most important regarding spatial encoding, the applicants reported recently having developed the MSVD, a powerful numerical mathematical tool that can determine near-optimal 3D spatial encoding. Second, a simple fast numerical procedure has been developed in their laboratory for the computation of RF pulse waveforms for implementing non-Fourier encodings using high flip angles ( 90=A1) for high SNR scans. Third, at their facility, they have the operating capability for near real-time adaptive 2D MRI using a commercial MR system with the minor modification of an additional attached workstation.