Technical Note Time-Resolved Contrast-Enhanced Carotid Imaging Using Undersampled Projection Reconstruction Acquisition Jiang Du, PhD, 1,2 * Sean B. Fain, PhD, 1 Frank R. Korosec, PhD, 1 Thomas M. Grist, MD, 1 and Charles A. Mistretta, PhD 1 Purpose: To investigate the utility of nonuniform angular spacing of projections in a three-dimensional (3D) hybrid undersampled projection reconstruction (PR) acquisition for contrast-enhanced (CE) time-resolved carotid imaging. Materials and Methods: Carotid CE magnetic resonance angiography (CE-MRA) was performed on seven healthy volunteers using a time-resolved hybrid sequence that combined undersampled PR acquisition in-plane and Car- tesian slice encoding through-plane. The undersampling streak artifact comes mainly from the superior–inferior (S/I) direction in carotid imaging and is suppressed by nonuniform distribution of the projections. Phantom and volunteer studies were performed to demonstrate its effi- cacy. Results: The undersampling streak artifact was signifi- cantly suppressed through a nonuniform distribution of the projection angles with more projections aligned along the S/I direction. The hybrid PR sequence combined with nonuniform distribution of the projection angles provided time-resolved images of the carotid arteries with high tem- poral resolution (two seconds per frame) and high spatial resolution (1.0  1.0  1.5 mm 3 ) simultaneously. Conclusion: High-resolution dynamic imaging of the ca- rotid arteries is feasible with the use of a hybrid under- sampled PR acquisition. Undersampling streak artifact can be suppressed significantly through nonuniform distribu- tion of the projections. Key Words: time-resolved MRI; undersampled projection reconstruction; carotid artery imaging; CE-MRA; non-uni- form distribution J. Magn. Reson. Imaging 2007;25:1093–1099. © 2007 Wiley-Liss, Inc. NONINVASIVE MAGNETIC RESONANCE IMAGING (MRI) of the carotid arteries is routinely performed with three-dimensional (3D) contrast-enhanced MR angiog- raphy (CE-MRA) (1–10). The key to achieving successful CE-MRA images is to acquire the central lines of k- space during peak arterial enhancement; however, this is challenging because of the rapid arterial-venous transit time and short duration of the arterial phase (six to 10 seconds) (6 – 8). One way to perform single 3D volume CE-MRA is to coordinate the acquisition of the central k-space data with the peak arterial enhance- ment using bolus timing techniques (4 – 6). Another ap- proach is to perform time-resolved imaging with rapid acquisition of multiple 3D volumes throughout the pas- sage of the contrast bolus (8 –10). The latter approach provides information regarding the anatomy and flow dynamics without the need for bolus timing, and there- fore is insensitive to interpatient variability in contrast material arrival and delayed vessel filling (8). A previous study implemented dynamic CE-MRA (DCE-MRA) using fast multiphasic acquisitions of 3D data sets and a combination of sensitivity and gradient encodings, which provided a temporal resolution of up to four to six seconds per frame (10). Another approach uses variable k-space sampling and/or temporal inter- polation (11). A temporal resolution of up to four sec- onds per frame has been obtained with the 3D time- resolved imaging of contrast kinetics (TRICKS) technique (8). However, temporal resolution is traded for spatial resolution in Cartesian TRICKS acquisitions. It has been shown that undersampled projection re- construction (PR) imaging preserves image spatial res- olution at the cost of high-frequency aliasing (12). The combination of undersampled PR in-plane and TRICKS interpolation through plane (PR-TRICKS) provides both high temporal resolution and high in-plane spatial res- olution. One can further improve the spatial resolution in the through-plane by acquiring high-frequency slice- encoding data during the steady state of the contrast material (PR-HyperTRICKS). The efficiency of this tech- nique has been demonstrated in the lower extremities (13). In this study we investigated the application of PR-HyperTRICKS to dynamic carotid imaging. We com- pared streak artifact reduction methods based on non- 1 Department of Medical Physics and Radiology, University of Wiscon- sin, Madison, Wisconsin, USA. 2 Department of Radiology, University of California–San Diego, San Di- ego, California, USA. *Address reprint requests to: J.D., Department of Radiology, University of California–San Diego, 200 West Arbor Drive, San Diego, CA 92103- 8756. E-mail: jiangdu@ucsd.edu Received June 2, 2006; Accepted Nvember 28, 2006 DOI 10.1002/jmri.20890 Published online 4 April 2007 in Wiley InterScience (www.interscience. wiley.com). JOURNAL OF MAGNETIC RESONANCE IMAGING 25:1093–1099 (2007) © 2007 Wiley-Liss, Inc. 1093