Phase Contrast Ultrashort TE: A More Reliable Technique for Measurement of High-Velocity Turbulent Stenotic Jets Kieran R. O’Brien, 1 * Saul G. Myerson, 2 Brett R. Cowan, 3 Alistair A. Young, 1,4 and Matthew D. Robson 2 Accurate measurement of peak velocity is critical to the as- sessment of patients with stenotic valvular disease. Conven- tional phase contrast (PC) methods for imaging high-velocity jets in aortic stenosis are susceptible to intravoxel dephasing signal loss, which can result in unreliable measurements. The most effective method for reducing intravoxel dephasing is to shorten the echo time (TE); however, the amount that TE can be shortened in conventional sequences is limited. A new se- quence incorporating velocity-dependent slice excitation and ultrashort TE (UTE) centric radial readout trajectories is pro- posed that reduces TE from 2.85 to 0.65 ms. In a high-velocity stenotic jet phantom, a conventional sequence had >5% flow error at a flow rate of only 400 mL/s (velocity >358 cm/s), whereas the PC-UTE showed excellent agreement (<5% error) at much higher flow rates (1080 mL/s, 965 cm/s). In vivo feasi- bility studies demonstrated that by measuring velocity over a shorter time the PC-UTE approach is more robust to intravoxel dephasing signal loss. It also has less inherent higher-order motion encoding. This sequence therefore demonstrates po- tential as a more robust method for measuring peak velocity and flow in high-velocity turbulent stenotic jets. Magn Reson Med 62:626 – 636, 2009. © 2009 Wiley-Liss, Inc. Key words: phase contrast; turbulent jets; high-velocity; ultra- short TE The routine clinical examination for assessment of the severity of aortic stenosis is Doppler ultrasound, and if the beam is correctly aligned, this technique can reliably mea- sure the peak velocity in the turbulent stenotic jet. In some cases, limited acoustic windows and other anatomical fac- tors can make this alignment difficult or result in poor image quality (1,2). For these patients an invasive exami- nation transesophageal echocardiography or catheteriza- tion is needed to clarify the result. Phase contrast (PC) magnetic resonance (MR) is an attractive noninvasive al- ternative that allows free positioning of the imaging plane and typically yields accurate measurements of velocity (3). Unfortunately, MR PC has been hindered by the question- able reliability of velocity measurements in moderate-se- vere turbulent stenotic jets in the aorta (4 – 8). Accurate classification of aortic stenosis severity is vital for the surveillance of stable patients, and timing of surgical in- tervention. When all the spins in a voxel have the same constant velocity PC will yield accurate results (9). As the range of velocities (and hence phase) increase in a voxel, signifi- cant signal loss can occur as a result of intravoxel dephas- ing, which results in unreliable measurements (10). Many different mechanisms have been found to influence intra- voxel dephasing signal loss in stenotic jets including high- er-order motion, e.g., acceleration (11–14), alignment of the image plane with the jet (15,16), voxel size and partial volume effects in voxels with large velocity distributions (10,17–20) the user-defined velocity encoding (Venc) (21), and turbulence imposing small velocity fluctuations su- perimposed on the principal velocity (22). Although it remains unclear which underlying mechanism is the most important, it is agreed that reducing the echo time (TE), defined as the time between the center of excitation and sampling the center of k-space, reduces intravoxel dephas- ing (13,20,23), improves signal to noise, and increases the reliability of PC velocity estimates. Gradient hardware improvements have led to stronger magnetic field gradients with higher slew rates, which can be utilized to shorten the TE. Previously we showed that in a high-velocity stenotic phantom shorter TE increased the signal intensity, accuracy, and reliability of velocity mea- surements (23). However, the ability to shorten TE in a conventional sequence using higher gradient performance is limited, and even with the most recent hardware, resid- ual errors are present at high flow-rates (23). In addition, the use of larger gradients and faster rise times results in larger errors from background phase due to eddy currents, thereby reducing some of the gains associated with im- proved gradient performance. It would therefore be ideal to reduce the TE beyond what is currently available in the literature (Table 1) without relying on higher gradient strengths and rise times. In MR angiography, ultrashort TE (UTE) approaches en- able very short TE acquisitions and centric-radial acquisi- tions, and variants used in UTE methods have been shown to improve image quality in regions of pulsatile flow (24). The mechanisms for this improvement are the oversam- pling of the center of k-space (25), and inherent minimi- zation of the first moment of the readout gradient provided by center-out k-space trajectories (26). PC sequences that employ spiral readout trajectories, another center-out-k- space trajectory, have been previously implemented (17); however, these have long readout trajectories and are thus sensitive to spin mixing and in-plane motion (20). 1 Bioengineering Institute, University of Auckland, Auckland, New Zealand. 2 Oxford University Centre for Clinical Magnetic Resonance Research (OCMR), University of Oxford, United Kingdom. 3 Centre for Advanced MRI, University of Auckland, Auckland, New Zealand. 4 Department of Anatomy with Radiology, University of Auckland, Auckland, New Zealand. Grant sponsor: New Zealand government Top Achievers scholarship; Grant sponsor: Dr. Joost Henkel Foundation stiftung; Grant sponsor: NZ Freema- sons; Grant sponsor: William Gerogetti Trust; Grant sponsor: Shritcliffe Fel- lowship (all to K.O’B.). *Correspondence to: Kieran O’Brien, Bioengineering Institute, Faculty of En- gineering, University of Auckland, Private Bag 92019, Auckland Mail Centre, Auckland 1142, New Zealand. E-mail: k.obrien@auckland.ac.nz Received 1 October 2008; revised 7 March 2009; accepted 26 March 2009. DOI 10.1002/mrm.22051 Published online 1 June 2009 in Wiley InterScience (www.interscience.wiley. com). Magnetic Resonance in Medicine 62:626 – 636 (2009) © 2009 Wiley-Liss, Inc. 626