Recent improvements in parallel-imaging performance in MRI have been driven by the use of greater numbers of radiofrequency (RF) surface coils placed in an array so as to maximize the ability to unwrap the aliasing that occurs along a linear phase encode direction when the data is undersampled (accelerated). This approach to acceleration is maturing and providing diminishing returns as more and more receiver coils are used, due to coil-coupling problems, which dominate as the size of the coil elements decreases and the number of elements increases. This proposal is aimed at providing further acceleration in parallel imaging through a reassessment of the magnetic field gradients used for spatial encoding. By considering the joint contributions of spatial encoding of the receiver coils and the magnetic fields, we can develop a more efficient approach to spatial encoding. A simple solution to this problem is to use a nonlinear magnetic field gradient, the Z2 gradient. Such a gradient shape (relative to linear X and Y gradients) provides spatial encoding more complementary to that provided by the receiver coils. The approach we have developed, O-space imaging - so named because the isofrequency contours during the readout are in the shape of concentric rings rather than columns as with a linear X-readout - was initially introduced by us in 2010. The theory and preliminary data generated using a small Z2-gradient insert on a clinical MRI scanner have provided evidence that substantial increases in acceleration can be achieved with modest numbers of receiver coils. This proposal is aimed at scaling these tests up to human imaging levels using a head-insert Z2-gradient coil fully integrated with the Siemens 3T Trio scanner. The O-space imaging approach is a general acceleration method that can be adapted to almost any pulse sequence. We will test, in phantom and human imaging experiments, the capabilities of O-space imaging to outperform conventional SENSE imaging using modified spin echo, fast spin echo and 3D imaging sequences at a range of resolutions and acceleration factors and using both an 8-channel and a 32-channel Tx/Rx coil. The end result of this work will be a demonstration of the viability of this new methodology for providing highly accelerated parallel imaging. The project is highly innovative and by reconsidering the spatial encoding gradients it opens up a new area of research in MR accelerated imaging. The project is significant in that providing an acceleration factor of 2 or more to a number of standard clinical MR pulse sequences could provide a huge benefit to public health, allowing for higher resolution, or more detailed examinations, and significantly increased throughput improving access to MRI and/or lowering per-scan costs.
Current methods in accelerating MRI image acquisitions have focused on receiver coil arrays with more and more elements to improve acceleration. This work represents a paradigm shift in parallel imaging by introducing a nonlinear Z2 gradient to produce spatial encoding complementary to the rf coil encoding increasing maximum achievable accelerations by more than a factor of 2. The project carries a natural benefit to the public healt through acceleration of the data acquisition process, allowing for higher resolution, more thorough examinations or shorter scan times to allow more patient throughput decreasing the PET scan cost of MRI.
|Kopanoglu, Emre; Constable, R Todd (2015) Radiofrequency pulse design using nonlinear gradient magnetic fields. Magn Reson Med 74:826-39|
|Li, Shu; Chan, Cheong; Stockmann, Jason P et al. (2015) Algebraic reconstruction technique for parallel imaging reconstruction of undersampled radial data: application to cardiac cine. Magn Reson Med 73:1643-53|
|Tam, Leo K; Galiana, Gigi; Stockmann, Jason P et al. (2015) Pseudo-random center placement O-space imaging for improved incoherence compressed sensing parallel MRI. Magn Reson Med 73:2212-24|