A flexible view ordering technique for high‐quality real‐time 2DFT MR fluoroscopy

Abstract

A method to tailor the view order to the reconstruction cycle is introduced for real-time MRI. It is well known that view sharing and oversampling central k-space views can improve the temporal resolution of gradient-echo pulse sequences. By ordering phase-encodes to synchronize k-space acquisition with the reconstruction cycle, apparent temporal resolution can match the frame rate with as few as one-fourth of the phase-encodes sampled per reconstruction. Spatial resolution is maintained by periodically updating high spatial frequencies. In addition to apparent temporal resolution, three other criteria for real-time imaging are identified and evaluated: display latency, dispersion, and frame-to-frame consistency. Latency is minimized by ordering views in a reverse-centric manner within each reconstruction interval, sampling high-energy views immediately prior to beginning reconstruction. Dispersion is kept low and consistent by synchronizing acquisition and reconstruction, thus avoiding poorly timed reconstruction instances. The ability to provide MR images to the operator in real time is becoming critical to more and more applications such as contrast bolus triggering (1), interventional MR (2-4), and cardiac imaging (5-7). Real-time imaging requires continuous, efficient data acquisition as well as rapid, on-line reconstruction and display.MR ''fluoroscopy'' was first demonstrated using a fast two-dimensional Fourier transform (2DFT) gradient-echo sequence with sequential phase-encode sampling (8). Using a view-sharing reconstruction, a single measurement of a given view may be used in the formation of more than one image; views are acquired continuously, but only the most recent measurement for each view is used in reconstructing a given frame, thus allowing more frequent image updating than if waiting for an entire set of new views to be collected. Similar view-sharing reconstruction techniques have subsequently been implemented for radial (9), spiral (10), and echoplanar imaging (EPI) (11) acquisitions. Although the data acquisition rate in 2DFT gradient-echo imaging is typically lower than for these other pulse sequences, 2DFT gradient-recalled echo(GRE) offers a number of distinct advantages for fluoroscopic imaging. These include robustness against off-resonance effects, negligible T2* resolution limitations, anti-alias filtering in the frequency-encode direction, moderate gradient strength requirements, relatively easy implementation, straightforward image reconstruction, and arbitrary ordering of the phase encodings.Because low spatial frequency views provide gross positional and contrast information while high-frequency views provide edge and detail information, temporal resolution is determined not only by the frequency of reconstruction, but also by the frequency at which the central region of k-space is sampled. Considering this, it was proposed in ref. 8 that temporal resolution could be increased by sampling the central region of k-space more frequently than the remainder. Several investigators h...

Steady‐state preparation for spoiled gradient echo imaging

“…Partial view updating (as described in Ref. 13) with 96 views acquired per image, was used to accelerate temporal resolution and reduce display latency. Whenever steady state was disturbed by changing the slice offset, orientation, or flip angle, k-space data buffers were cleared (to avoid artifacts due to partial updating) and steady-state preparation was immediately performed prior to resuming imaging.…”

Temporal frequency analysis of dynamic MRI techniques

“…Several sampling strategies have been proposed to improve both the temporal and spatial resolution. Such methods include keyhole imaging (10 -13), reduced-encoding imaging through generalized-series reconstruction (RIGR) (14 -16), time-resolved imaging of contrast kinetics (TRICKS) (9,17,18), and flexible view ordering technique for realtime 2DFT MR fluoroscopy (19). The goal behind all of these methods is to reduce the amount of data needed for a given spatial and temporal resolution.…”