FULL PAPER Fat Quantification Using an Interleaved Bipolar Acquisition Abraam S. Soliman, 1,2 * Curtis N. Wiens, 3 Trevor P. Wade, 2,4 and Charles A. McKenzie 1,2,4 Purpose: To demonstrate a new multigradient echo bipolar acquisition sequence for fat quantification. Theory and Methods: A multiecho bipolar acquisition is used such that the even echoes have opposite polarity to the odd echoes. In addition, the readout gradients alternate their polar- ities every other phase-encode line. Each echo, therefore, con- sists of phase-encode lines with both positive and negative polarities. Phase-encodes acquired with the same polarity are grouped together, and parallel imaging reconstruction is used to obtain two full k-space maps with opposite readout polar- ities at all the echoes. By complex averaging, the inconsistent phase errors between odd and even echoes are removed and water/fat separation techniques used with conventional unipo- lar sequences can be performed. Results: Phantoms and in vivo experiments demonstrated accurate fat fraction and increased signal to noise ratio effi- ciency compared with the established unipolar acquisition. Artefacts due to phase and magnitude errors of bipolar acqui- sition were eliminated in all experiments. Conclusion: The interleaved bipolar sequence is an efficient technique for fat quantification. It demonstrated accurate fat measurements in a shorter scan time compared with the standard unipolar sequence. Magn Reson Med 75:2000– 2008, 2016. V C 2015 Wiley Periodicals, Inc. Key words: chemical-shift imaging; bipolar acquisition; fat quantification; IDEAL imaging INTRODUCTION Multigradient echo sequences have been widely used for chemical-shift encoded imaging, particularly water/fat separation. Images are acquired at several echo times (TEs), where fat and water exhibit different phase shifts (1,2). The acquired images are subsequently postpro- cessed to obtain water and fat components (3–9). To achieve an accurate fat measurement, several confound- ing factors must be considered during the reconstruction process, particularly B0 magnetic field inhomogeneities (1,2), T 2 * decay (10), T 1 -bias (11), noise-related bias (11), eddy-currents (12,13), temperature bias (14), and accu- rate spectral modeling of fat (15). Typically, six echo times are recommended for accurate fat measurement (10,15,16). On the acquisition side, parameters such as the flip angle, first echo time (TE 1 ), and echo spacing also influ- ence the signal to noise ratio (SNR) performance of the quantification process. Low flip angles are usually rec- ommended to reduce T 1 -bias (11). The echo combination (TE 1 and echo spacing) are kept as short as possible to improve SNR (12,17). Selection of TE 1 is restricted by the maximum gradient strength, the maximum achieva- ble slew-rate during the readout prewinder gradient and the time required for the RF pulse. Echo spacing varies with the selection of receiver bandwidth, spatial resolu- tion, field of view (FOV), and gradient performance (slew rate and maximum gradient strength). It is impor- tant to choose an echo spacing that provides the best SNR performance (12,17). In addition, short echo spacing is desired to increase the range of frequency offsets in which water and fat components can be unambiguously distinguished during the reconstruction process (2). Multiecho acquisition can be achieved using either uni- polar or bipolar readout gradients. In a unipolar acquisi- tion, the echoes have the same readout polarity while in a bipolar acquisition the echoes are acquired with both pos- itive and negative readout gradient polarities. Typically, fat quantification is performed using six unipolar readout gradients separated by flyback gradients (10,15,16). These long flyback gradients increase the minimum achievable echo spacing, heightening the ambiguity in identifying fat and water species. To achieve optimal echo spacing, the echoes are often acquired in an interleaved manner with multiple shots. For instance, the six echoes are acquired over 6/n repetition times (TRs) (i.e., over 6/n shots), where n is the number of echoes per TR. This also increases the acquisition time. In a bipolar acquisition the flyback gradients are omitted and the six echoes alternate their polarities (18). By removing the flyback gradients, shorter echo spacing can be achieved, often allowing the acquisition of all of the echoes at optimal echoes spacing in one repetition time (i.e., a single shot). Consequently, a shorter scan time can be achieved. Therefore, bipolar acquisitions have higher SNR efficiency, as more of the scan time is spent collecting the data, where SNR effi- ciency is defined as the SNR normalized by the acquisi- tion time (SNR / sqrt (acquisition time)). However, several challenges accompany bipolar sequences (18,19). First, phase errors induced by readout gradient delays and eddy-currents can severely distort the reconstruction if not explicitly accounted for. Although these phase errors also exist in unipolar 1 Biomedical Engineering, Western University, London, Canada. 2 Robarts Research Institute, Western University, London, Canada. 3 Department of Radiology, University of Wisconsin-Madison, Madison, Wisconsin, USA. 4 Department of Medical Biophysics, Western University, London, Canada. *Correspondence to: Abraam Soliman, Ph.D., Biomedical Engineering Graduate Program, Natural Sciences Building, Room 9, Western University, 1151 Richmond Street, London Ontario Canada N6A 5B7. E-mail: asoliman@alumni.uwo.ca Received 24 March 2015; revised 20 May 2015; accepted 21 May 2015 DOI 10.1002/mrm.25807 Published online 20 June 2015 in Wiley Online Library (wileyonlinelibrary. com). Magnetic Resonance in Medicine 75:2000–2008 (2016) V C 2015 Wiley Periodicals, Inc. 2000