Synopsis

Keywords: Cardiovascular: Cardiac, Cardiovascular: Cardiac function, Image acquisition: Motion Correction

Non-rectilinear signal readouts offer both unprecedented challenges and opportunities both on the acquisition and on the reconstruction side. These will be discussed with a particular focus on cardiovascular applications. In this context, fat suppression strategies, 2D imaging, 3D imaging and even excitation k-space will be discussed.

While “Off-the-Grid” commonly refers to “not using or depending on public utilities, especially the supply of electricity”, here it refers to non-uniformly spaced signal sample points that do not necessarily align with a Cartesian grid of discrete k-space positions. In some sense, and since Fourier data are not measured on a regularly spaced rectlinear grid in that case, it also prevents us from directly using the “public utility” fast Fourier transform (FFT). Therefore, extra efforts including but not limited sampling density compensation, where data are normalized by the inverse of their sampling density, and followed by the interpolation of the data onto a Cartesian grid (“gridding”) using a convolution kernel, are needed prior to fast Fourier transforming the k-space data into the image domain (non-uniform FFT or NUFFT) (1). However, this also begs the question of what the benefits of non-rectilinear sampling are that outweigh the expense of such additional efforts. Advantages may include scan efficiency improvements in terms of k- space coverage per unit time, robustness against motion secondary to an averaging effect of low spatial frequencies, the ability to minimize the effects of T2* decay in the center of k-space, and general flexibility in measurement strategies. Interestingly, and with very few exceptions, there is a general lack of non-Cartesian signal sampling strategies being used in daily routine clinical practice. Hindrances include sensitivity to off-resonance and resultant blurring, more stringent requirements for shimming, and sometimes the limited effectiveness of preparatory pulses to generate contrast. In this context, fat saturation poses a particularly interesting challenge for increased contrast on the one hand and to help minimize off-resonance blurring on the other hand. Spectral-spatial radiofrequency (RF) pulses enable water-only excitation at the expense of the RF pulse duration (2). While spectral-spatial RF excitations would theoretically be effective for most Cartesian signal-readouts, their use unfortunately tips the balance of the duty cycle (how much time is spent for RF excitation relative to signal sampling) in an unfavorable direction and prolongs scanning time when compared to a simple fat saturation pre-pulse concept that lends itself well for fat suppression in Cartesian imaging. Conversely, in the case of a spiral signal-readout which is typically prolonged, this balance may tip in favor of a spectral-spatial excitation and provides an extremely powerful means to help minimize off-resonance artifacts on the one hand and enhance fat-to-tissue contrast on the other (2). However, with a radial signal-readout that is also non-Cartesian, spectral-spatial RF excitations affect the duty cycle and thus prolong scanning times. Compromises include spectral-only excitations where slice- or volume-selectivity can be avoided because of large-volume imaging that is enabled through intrinsic signal oversampling in every spatial direction for radial imaging (3). For radial imaging, fat saturation using a pre-pulse concept rather than a spectral excitation is possible, but its effectiveness may be affected by the re-growth of the longitudinal fat magnetization during the series of signal-readouts. Note that this problem is accentuated for radial imaging where every single readout traverses the center of k-space. An interesting hybrid implementation of Cartesian and non-Cartesian imaging includes the “Propeller” technique (4) where data are collected in concentric rectangular strips rotated about the center of k-space. Of such a strip of parallel k-space readouts, only one such readout goes through the center of k-space per angular increment. This uniquely combines the advantages of Cartesian and non-Cartesian techniques and pre-pulse concepts including fat saturation may become more effective than for pure radial imaging. Regardless, non-Cartesian imaging strategies offer unique opportunities, particularly in the domain of cardiovascular imaging. This begins with the seminal and initial report on spiral coronary artery imaging by Craig Meier in 1990 (2) and various 3D variants that were introduced later (5). Nowadays, free-running 3D radial imaging with k-space segments rotated by the golden angle (6) has helped shift a paradigm in cardiovascular MR where painstaking and labor-intensive prospective scan parameter adjustment has been replaced with a highly flexible a posteriori extraction of the desired anatomical or functional information (7). Simultaneously, spiral imaging has critically enabled MR fingerprinting (8), another landmark in MR methods development. And these latter two examples have particularly benefitted from the ever-increasing computing power since modeling, dictionary fitting, and compressed sensing reconstruction (9) can now be exploited to leverage the full potential of these non-Cartesian approaches. With the advent of commercial low-field systems where both T2* and B0 inhomogeneities are reduced, the breadth of applications that benefit from non-Cartesian signal sampling is likely to be further expanded. During my presentation, examples mostly focused on cardiovascular imaging will be presented to document the wealth of possibility provided by going “Off-the-Grid”. At this juncture, I would be amiss not to mention briefly that non-Cartesian trajectories not only offer opportunity in “signal-readout k-space” but also in “excitation k-space” (10) and a few examples will be discussed as well. “Off-the-Grid” also means being disconnected from the daily grind – and the ISMRM in Toronto will be the perfect opportunity to enjoy that and to network with other “Off-the-Gridders” as it were. Looking forward to seeing you in Toronto, one of the most multicultural and cosmopolitan cities in the world!

Acknowledgements

No acknowledgement found.