Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 1.
Published in final edited form as: Magn Reson Imaging. 2020 Oct 29;75:156–161. doi: 10.1016/j.mri.2020.10.015

A Novel Spectrally Selective Fat Saturation Pulse Design with Robustness to B0 and B1 Inhomogeneities: A Demonstration on 3D T1-weighted Breast MRI at 3T

Feng Xu 1,2,†,*, Wenbo Li 1,2,†, Dapeng Liu 1,2, Dan Zhu 3, Michael Schär 1, Kelly Myers 1, Qin Qin 1,2
PMCID: PMC7683379  NIHMSID: NIHMS1642196  PMID: 33130057

Abstract

Purpose:

Spectrally selective fat saturation (FatSat) sequence is commonly used to suppress signal from adipose tissue. Conventional SINC-shaped pulses are sensitive to B0 off-resonance and B1+ offset. Uniform fat saturation with large spatial coverage is especially challenging for the body and breast MRI. The aim of this study is to develop spectrally selective FatSat pulses that offer more immunity to B0/B1+ field inhomogeneities than SINC pulses and evaluate them in bilateral breast imaging at 3T.

Materials and Methods:

Optimized composite pulses (OCP) were designed based on the optimal control theory with robustness to a targeted B0/ B1+ conditions. OCP pulses also allows flexible flip angles to meet different requirements. Comparisons with the vendor-provided SINC pulses were conducted by numerical simulation and in vivo scans using a 3D T1-weighted (T1w) gradient-echo (GRE) sequence with coverage of the whole-breast.

Results:

Simulation revealed that OCP pulses yielded almost half of the transition band and much less sensitivity to B1+ inhomogeneity compared to SINC pulses with B0 off-resonance within ±200 Hz and B1+ scale error within ±0.3 (P<0.001). Across five normal subjects, OCP FatSat pulses produced 25–41% lower residual fat signals (P < 0.05) with 27–36% less spatial variation (P<0.05) than SINC.

Conclusion:

In contrast to conventional SINC-shaped pulses, the newly designed OCP FatSat pulses mitigated challenges of wide range of B0/ B1+ field inhomogeneities and achieved more uniform fat suppression in bilateral breast T1w imaging at 3T.

Keywords: Spectrally selective RF pulse, fat saturation, fat suppression, B0 insensitive, B1+ insensitive, optimal control, breast MRI

1. Introduction

A number of fat suppression or water-fat separation techniques are routinely used in various clinical MRI protocols to mitigate the signal or artifacts arise from adipose tissue due to its short T1 relaxation or different chemical shift [1,2]. Available methods include short tau inversion recovery (STIR) [3], chemical shift selective fat saturation (FatSat) [4,5], or fat inversion [6], chemical shift based water-fat separation using Dixon [7–14], and water-selective excitation [15–17].

The conceptually straightforward FatSat module applies a spectrally selective excitation pulse to flip only the lipid signal onto the transverse plane with a following spoiler gradient and can be added as a pre-pulse to most pulse sequences. Typically, a frequency-selective RF pulse is applied to keep the main fat peaks in the saturation band and the water frequency in the pass band. However, the conventional SINC-shaped RF pulses commonly adopted by FatSat yield a relatively broad transition band, which could perturb the water signal or generate less nutation of lipid signal in the presence of B0 off-resonance. More importantly, flip angle of the saturation band offered by a SINC pulse is scaled by B1+ errors. Thus its sensitivity to the B0 and B1+ inhomogeneities limits the performance of FatSat at high field or with a large spatial coverage. Uniform fat saturation is especially challenging for breast MRI at 3T [18–21] with heterogeneous distribution of B0 field (standard deviation up to 60–70 Hz [22]) and B1+ field (left-to-right ratio up to 1.2–1.3 [23,24]) over bilateral breast regions. Clinically, diagnosing a small enhancing breast tumor would benefit from better fat suppression.

Optimal control theory was introduced into RF pulse design in NMR [25,26]. It provides the flexibility for systematically imposing desirable constraints on the spin system’s evolution. Optimized composite (OCP) pulses based on optimal control algorithms have been shown to improve the broadband excitation and inversion, as well as bandwidth selectivity [25–28]. Low-flip-angle OCP pulses achieved uniform excitations in wide B0/B1+ ranges for both anatomical and angiography sequences [29–31].

This work aimed to demonstrate a novel design of FatSat pulses based on the optimal control theory with robustness to a targeted wide range of B0/B1+ conditions. The performance of OCP FatSat was compared with the vendor-provided Sinc-shaped FatSat pulse using both numerical simulation and in vivo scans applying T1-weighted (T1w) gradient-echo (GRE) sequences with whole-breast coverage on normal volunteers at 3T.

2. Methods

2.1. Multi-peak fat model based optimal flip angles for FatSat pulses

For spectrally selective fat suppression, it is often assumed that fat has a single resonance frequency at about 3.4 ppm downfield from water. However, many MR spectroscopy studies revealed that fat has a complexed multi-peak spectrum [32–35]. Our simulation adopted a human fat model with five lipid peaks at 5.3, 4.2, 2.1, 1.3, 0.9 ppm and the corresponding percentages of total fat signal were 5%, 4%, 12%, 70%, 9%, respectively [32,33]. Given that the water peak is at 4.7 ppm, the fat peaks at 2.1 ppm, 1.3 ppm, 0.9 ppm with a total of 91% of fat signal can be saturated by the FatSat pulses. In contrast, the two smaller lipid peaks (5.3 ppm, 4.2 ppm) surrounding the water peak and accounting for 9% of total fat signal are hardly affected by FatSat pulses. To further mitigate these near-water fat peaks, the flip angles of FatSat pulses can be adjusted so that fat peaks buried under the water peaks could be more or less canceled by the residual of main fat peak [36,37]. The optimization strategy depends on TE and TR. TEs at which the water and the main fat peak (3.4 ppm chemical shift) form in-phase or opposed-phase conditions are considered: 95° and 85° FatSat pulses (cos(95°) = −0.09 and cos(85°) = 0.09) on the main fat peak would be desired to cancel the 9% fat signal near the water resonance at in-phase and opposed-phase TEs, respectively. Since fat peaks all have relatively short T1 values [36,38], full recovery of all fat signals were assumed in this simulation.

2.2. Designing OCP pulses for FatSat

OCP was designed for B0 and B1+ insensitive excitation through an optimal control routine in the MatPulse toolbox on Matlab (Mathworks) [28,30,39]. Here we designed saturation pulses which were composed of hundreds of block-shaped subpulses with their amplitudes and phases derived through numerical optimization. The desired range of immunity to the B0/B1+ offset occurred in the breast at 3T is: B0 = ±200 Hz and B1+ scale = ±0.3 (estimated from our experiments, see Supporting Figure S1). The B1+ amplitude was set to 13 μT, slightly lower than the peak amplitude of 13.5 μT of the internal whole-body transmit coil of our 3T scanner. Other optimizing conditions included the saturation bandwidth = 720 Hz and the transition bandwidth = 75 Hz. Initial pulses used the Shinnar–Le Roux (SLR) algorithm to generate the respective RF pulses with corresponding flip angle, pulse length, dwell time and frequency response. Only a single side of the profile was optimized by the optimal control routine in order to reduce the pulse duration [28]. Numerical simulations of the Bloch equations were performed using Matlab to examine the responses of the longitudinal magnetizations (Mz) following the FatSat pulses under various B0/B1+ conditions. T1 and T2 effects were not accounted for in this simulation. Comparison were first made between a 12 ms 90° OCP pulse with 600 subpulses and a vendor-provided FatSat pulse, a 7.5 ms 90° SINC-shaped pulse with windowing (labeled as SINC in this work for simplicity). OCP pulses with 95° and 85° flip angles with consideration of the multi-peak fat model were also designed for in vivo evaluations in comparison to single-peak derived 90° OCP pulse.

2.3. Experiments

2.3.1. In vivo experiments

Non-contrast-enhanced experiments were performed on a 3T Philips Ingenia scanner using the body coil for RF transmission (maximum B1+ amplitude 13.5 μT) and a 16-channel breast coil for signal reception. Five healthy female volunteers (45±15 years, range 27–61 years) were enrolled. Each participant provided written informed consent that was approved by the Johns Hopkins Medicine Institutional Review Board. Participants were scanned with prone and feet-first position.

FatSat modules were compared between OCP and SINC pulses using in-phase TE (= 2.30 ms) and opposed-phase TE (= 1.15 ms). SINC 90°, 95°, and 105° were performed at both TEs. SINC 85° was performed at opposed-phase TE on four subjects in addition. OCP 90° and OCP 95° were performed at in-phase TE; while OCP 90° and 85° were performed at opposed-phase TE. The FatSat pulses were immediately followed by a spoiling gradient to de-phase the fat signal in the transverse plane. For reference, no FatSat scans were also conducted by turning the FatSat modules off.

The 3D T1w GRE acquisition with whole-breast coverage applied the multi-shot turbo field echo (TFE). An axial slab was acquired with left-right phase-encoding and the following parameters: FOV = 200 (Left to Right) x 366 (Anterior to Posterior) x 120 (Foot to Head) mm3, acquired resolution = 1.0 × 1.0 × 2.0 mm3, reconstructed resolution = 0.85 × 0.85 × 1.0 mm3, TR/TE = 3.9/2.3 ms (in-phase) and 2.8/1.15 ms (opposed-phase), flip angle = 7°, low-high profile ordering, TFE factor = 70, SENSE factor = 2.5 × 1.5, shot number = 91, TFE acquisition window = 276 and 196 ms, respectively, for the two types of TEs, TRshot = 500 ms, and total scan duration was 0.8 min for each scan with or without FatSat.

B0 maps were obtained by a standard 3D GRE sequences with two water-fat in-phase TE (2.3 ms and 4.6 ms). FOV had the same coverage as the T1w 3D GRE sequence, acquisition resolution was 3×3×4 mm3, TR = 9.5 ms, flip angle = 30°. B1+ maps were measured with the DREAM method [40,41], using multi-slice 2D acquisition, TR extension = 1500 ms, TR/TE1/TE2 = 5.5/2.3/2.9 ms, flip angle = 15°, STEAM angle = 40°. FOV and acquisition voxel size were implemented with the same specifications as the B0 mapping. All these scans used vendor-provided fully automated shimming methods called “smart” for both B0 and B1+ shimming, which considers geometries of individual breasts.

Lastly, a six-point 3D GRE Dixon [34] was used to obtain separated water/fat and a fat fraction map. Dixon parameters were TR/TE/echo spacing = 8.1/1.33/1.1 ms, flip angle = 3°, acquired resolution = 1.5 × 1.5 × 2.0 mm2, SENSE factor = 2.0 × 1.0, with the same coverage as the T1w GRE sequence. The scan duration was 1 min.

2.3.2. Data analysis

Whole-breast mask was manually outlined for each participant. Fat masks were determined by thresholding larger than 80% of the fat fraction maps obtained by Dixon scans. Two metrics were compared across FatSat pulses within the breast fat masks: the spatial mean and standard deviation (STD) of residual fat signal with FatSat, both divided by the averaged fat signal without FatSat. The lower the normalized mean and STD, the higher the efficacy and uniformity of the fat suppression, respectively. Two-tailed paired t-test with equal variance were used for testing any difference between the best OCP and the best SINC scans. P value < 0.05 was considered statistically significant.

3. Results

3.1. Simulation

Figure 1 displays the amplitude and phase modulations of SINC 90° and OCP 90° and their Mz responses over the B0 off-resonance from −700 Hz to 300 Hz at B1+ scale = 1 and the B1+ scale from 0.7 to 1.3. Note that the square of the amplitude of the RF pulse, which is proportional to the specific absorption rate (SAR), of the 12 ms OCP was about 66 times of the value of the single 7.5 ms SINC. Compared to SINC 90°, OCP 90° pulse offered less than half of the transition width, 175 Hz vs 65 Hz, which were determined by changes from 95% to 5% of equilibrium Mz (shown in brown color band in Figure 1 (b) and (e)). OCP 90° pulse is prominently less sensitive to B1+ inhomogeneity on the fat suppression band (from −700 Hz to −200 Hz) while preserving the water signal on the passband (from −200 Hz to 300 Hz) (Figure 1 (c) and (f)).

Figure 1.

Figure 1.

Open in a new tab

The pulse shapes of (a) SINC 90° and (d) OCP 90°; (b) and (e) are the simulated FatSat profiles with the longitudinal magnetization (Mz) as a function of B0 off-resonance with ideal B1+ scale = 1; (c) and (f) are FatSat profiles as functions of both B0 off-resonance and B1+ scale errors. The two vertical dot lines in (b,c,e,f) are the relative resonance frequencies of water (0 Hz) and the main fat peak (−443 Hz) at 3T. The brown vertical bands in (c) and (f) indicate the transition bands determined by changes from 95% to 5% of the equilibrium Mz. Compared to SINC 90°, OCP 90° pulse yielded almost half of the transition band (175 Hz vs. 65Hz) and much less sensitivity to B1+ inhomogeneity.

3.2. In vivo experiments

The measured B0 off-resonance and B1+ scale maps across the breasts of five subjects were analyzed. A combined 2D histogram displays the B0 and B1+ distribution from all voxels of whole breasts of all five subjects (Supporting Information Figure S1). Region inside the red box ranges B0 off-resonance within ±200 Hz and B1+ scale error within ±0.3, which contained 96% of all scattered data.

The performance of vendor-provided SINC and our proposed OCP pulses with different flip angles are compared in Figure 2 for in-phase TE for one subject in 50 s and in Figure 3 for opposed-phase TE for another subject in 40 s. These two subjects both presented relatively mild B0 inhomogeneities within each breast (a) but strong B1+ heterogeneity between left and right breasts (b). The fat signal adjacent to breast tissue appeared to be better suppressed in opposed-phase images than in-phase images, most likely due to the additional cancelation by the water signal at opposed-phase TE. Conventional SINC pulses manifested partial unsaturated fat signal at different locations (orange arrows) on both sides for each of the FatSat angles from 85° to 105° (f-h), indicating their sensitivity to B1+ inhomogeneities. In contrast, fat signal appeared more homogeneously suppressed using OCP pulses as expected. Compared to OCP 90° (i), OCP 95° (Figure 2j) and OCP 85° (Figure 3j) yielded better fat suppression at in-phase TE (Figure 2) and opposed-phase TE (Figure 3), respectively. Purity, uniformity, and contrast of fat-suppressed images in (j) were comparable to these qualities of Dixon water-only images in (e).

Figure 2.

Figure 2.

Open in a new tab

Comparing FatSat methods for in-phase TE on one healthy volunteer’s breast MRI. Both axial and sagittal views are displayed: (a) B0 and (b) B1+ maps; (c) 3D T1w GRE without FatSat; (d) Dixon fat-only image; (e) Dixon water-only image; 3D T1w GRE with (f) SINC 90°, (g) SINC 95°, (h) SINC 105°, (i) OCP 90° and (j) OCP 95°. SINC FatSat pulses suffered from sensitivity to B1+ inhomogeneities, as indicated by arrowheads. In comparison, fat signal appeared more homogeneously suppressed using OCP pulses. OCP 95° yielded more complete fat suppression than OCP 90° for the in-phase TE.

Figure 3.

Figure 3.

Open in a new tab

Comparing FatSat methods for opposed-phase TE on another healthy volunteer’s breasts. Both axial and sagittal views are displayed: (a) B0 and (b) B1+ maps; (c) 3D T1w GRE without FatSat; (d) Dixon fat-only image; (e) Dixon water-only image; 3D T1w GRE with (f) SINC 85°, (g) SINC 90°, (h) SINC 95°, (i) OCP 90° and (j) OCP 85°. SINC FatSat pulses suffered from sensitivity to B1+ inhomogeneities, as indicated by arrowheads. In comparison, fat signal appeared more homogeneously suppressed using OCP pulses. OCP 85° yielded more complete fat suppression than OCP 90° for the opposed-phase TE.

The relative ratios of the spatial mean of residual fat signal after different FatSat pulses to the averaged fat signal without FatSat within the individual fat mask are exhibited for TEs with in-phase (Figure 4a) and opposed-phase (Figure 4b) of each subject. Error bar indicates the normalized standard deviation (STD) of residual fat from the individual fat mask. All five subjects showed the same trend across the group, that is, multi-peak fat model based OCP 95° (in-phase) / OCP 85° (opposed-phase) both performed better than single-peak based OCP 90°, with a mean ratio of remaining fat signal, 0.10 vs. 0.18, and 0.15 vs. 0.27, respectively. Compared to SINC 105° and 85°, which were the best among SINC pulses for in-phase and opposed-phase, OCP 95° and 85° further reduced residual fat by about 41% (P = 0.003, 0.10 vs. 0.17) and 25% (P = 0.02, 0.15 vs. 0.20) and decreased remaining fat STD by 36% (P = 0.002, 0.07 vs. 0.11) and 27% (P = 0.0006, 0.11 vs. 0.15), respectively.

Figure 4.

Figure 4.

Open in a new tab

Residual fat fraction was obtained from the relative ratios of the spatial mean of residual fat signal after different FatSat pulses to the averaged fat signal without FatSat within the individual fat mask are exhibited for TEs with in-phase (a) and opposed-phase (b) of each subject. Error bar indicates the normalized standard deviation (STD) of residual fat from the individual fat mask. Compared to SINC 105°, the best SINC pulses for in-phase (a), OCP 95° further reduced residual fat by about 41% (P = 0.003, 0.10 vs. 0.17), and decreased remaining fat STD by 36% (P = 0.002, 0.07 vs. 0.11); Compared to SINC 85°, the best SINC pulses for opposed-phase (b), OCP 85° further reduced residual fat by about 25% (P = 0.02, 0.15 vs. 0.20), and decreased remaining fat STD by 27% (P = 0.0006, 0.11 vs. 0.15).

4. Discussion

In this study, we developed spectrally selective OCP FatSat pulses to uniformly suppress fat signal with little spatial dependence due to B0/B1+ distributions. Compared to SINC pulses, these novel OCPs expressed high efficacy and uniformity of fat suppression in both in-phase and opposed-phase TEs across whole breasts of several healthy volunteers.

Some strategies exist for conventional SINC pulses to be less affected by B0 or B1+ field inhomogeneities separately. To mitigate the susceptibility to B0 off-resonance, SINC pulses can be lengthened to reduce the transition band. For instance, stretching the 7.5 ms SINC 90° pulse one third longer to 12 ms curtails the transition band to half, 175 Hz vs. 107 Hz (Supporting Information Figure S2 a,b). Longer SINC pulses are commonly used at 1.5T as the chemical shift between water and fat peaks are halved compared to 3T. If the variation of B1+ scales is large, the uneven saturation remains problematic. To alleviate the sensitivity to B1+ field inhomogeneity, repeating the SINC 90° once or twice have been proposed, with prior studies demonstrating improved fat suppression using this technique for abdominal and breast imaging at 1.5T [42,43]. Our simulation showed that, within the saturation band, double SINC still exhibited some degree of B1+ dependence and triple SINC offered much improved immunity to B1+ variation (Supporting Information Figure S2 c,d). However, double or triple the 7.5 ms SINC 90° pulse only reduced the transition band to 154 Hz and 134 Hz, respectively. In contrast, the OCP pulses yielded superior results in both the saturation and transition bands. The 12 ms OCP 90°’s transition width was 65 Hz, which could be further reduced to 39 Hz when it was stretched to 19 ms (Supporting Information Figure S2 e,f). Furthermore, the end nutation angles after double or triple SINC pulses could only be close to 90°, and yet the multi-peak fat model based 95° or 85° would be desired to cancel the ~9% fat peaks near the water resonance at in- or opposed-phase TEs. It is worth noting that, when the pulse duration is prolonged, T2 relaxation during the RF pulses would hamper its performance. Taking T2 value of fat or breast tissue at 3T as 53 ms [38], simulation of the 12 ms OCP pulse showed a slight signal reduction (2~4%) near the transition band (data not shown).

One disadvantage of the OCP pulse is its significantly higher specific absorption rate (SAR) which is proportional to the square of the amplitude of the RF pulse. For our FatSat prepared T1w scans with 500 ms shot intervals, the exerted total SAR was 71% of the torso SAR limit for the OCP compared to the 7% for the single SINC FatSat. This would restrict its utilization to certain SAR-conscious pulse sequences. As other conventional frequency-selective RF pulses, the SAR of the current OCP pulse shape can be lowered proportionally, by rescaling and stretching to longer pulse duration, but at the cost of narrower bandwidth. OCP pulses with 10 ms, 8 ms, 6 ms, and 4 ms pulse duration were also designed using the same procedure and compared with the 12 ms OCP pulses on some volunteers. Shorter OCP pulses lowered SAR proportionally, but with less uniform saturation band and less sharp transition band, which yielded slightly uneven performance of fat suppression in vivo (Supporting Information Figure S3). Minimizing SAR as an additional constraint in the optimal control routine [28] would be desired in future investigations.

As discussed in [30], the optimization routine of MatPulse used in this study did not ensure global optimization. Four different initiating pulses (minimal-phase, maximum-phase, linear-phase SLR pulses, and the vendor-provided SINC pulse shape) were compared and the results were shown in Supporting Information Figure S4. As can be seen, the derived OCP pulses depended on the initiating pulses and did not converge to a single pulse shape, but their Mz-responses in the wide B0/B1+ ranges did not differ much between each other. Adiabatic pulses such as hyperbolic secant shape could be tested for saturation pulses [28]. For each run of the OCP search algorithm for each input, it took up to 2 days on a desktop equipped with an 8-core 4.7GHz CPU. With pulse duration fixed, more subpulses could be used for better performance, which would require longer optimization process with increased degrees of freedom. At this stage, designing OCP is a more heuristic process than systematical or analytical. The waveforms of OCP 90°, 95°, and 85° pulses are provided in a supporting text file named OCP90_95_85_12ms_600points_FatSat_pulse_waveforms.txt.

In addition to the T1-weighed images, the OCP FatSat pulses could be employed during contrast preparation for many novel pulse sequences, such as T2-weighted [44,45], diffusion [46,47] and perfusion weighted MRI [48,49] or MR angiography [29,50]. They should perform well in other body parts (i.e., cardiac, spine, head/neck, foot/ankle) at 3T within the targeted B0/B1+ ranges. Alternatively, for accurate water-fat separation and fat quantification, Dixon-based methods would need to acquire 3–6 echoes for correcting B0 field inhomogeneity [9] and employing multi-peak fat model [34], thus demanding relatively longer time for both acquisition and reconstruction [2].

5. Conclusion

New spectrally selective fat saturation OCP pulses were designed based on optimal control theory for a targeted wide range of B0 off-resonance and B1+ scales. When compared to conventional SINC shaped pulses, these OCP pulses resulted in significantly more robust fat suppression for breast MRI when applied on a 3D T1w sequence with whole breast coverage at 3T.

Supplementary Material

1

Acknowledgments

Grant support from NIH R01 HL138182, NIH R01 HL144751, NIH S10 OD021648.

Footnotes

Prepared for submission as a Technique Note in Magnetic Resonance Imaging (MRI)

Declarations of interest: none

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • [1].Delfaut EM, Beltran J, Johnson G, Rousseau J, Marchandise X, Cotten A. Fat suppression in MR imaging: Techniques and pitfalls. Radiographics 1999;19:373–82. 10.1148/radiographics.19.2.g99mr03373. [DOI] [PubMed] [Google Scholar]
  • [2].Bley TA, Wieben O, François CJ, Brittain JH, Reeder SB. Fat and water magnetic resonance imaging. J Magn Reson Imaging 2010;31:4–18. 10.1002/jmri.21895. [DOI] [PubMed] [Google Scholar]
  • [3].Bydder GM, Pennock JM, Steiner RE, Khenia S, Payne JA, Young IR. The short TI inversion recovery sequence-An approach to MR imaging of the abdomen. Magn Reson Imaging 1985;3:251–4. 10.1016/0730-725X(85)90354-6. [DOI] [PubMed] [Google Scholar]
  • [4].Haase A, Frahm J, Hanicke W, Matthaei D. 1H NMR chemical shift selective (CHESS) imaging. Phys Med Biol 1985;30:341–4. 10.1088/0031-9155/30/4/008. [DOI] [PubMed] [Google Scholar]
  • [5].Frahm J, Haase A, Haenicke W, Diet M, Bomsdorf H, Helzel T. Chemical Shift Selective MR Imaging Using a Whole-Body Magnet. Radiology 1985;156:441–4. [DOI] [PubMed] [Google Scholar]
  • [6].Kaldoudi E, Williams S, Barker G, Tofts P. A chemical shift selective inversion recovery sequence for fat-suppressed MRI: theory and experimental validation. Magn Reson Imaging 1993;11:341–55. [DOI] [PubMed] [Google Scholar]
  • [7].Dixon, Thomas W. Simple Proton Spectroscopic Imaging. Radiology 1984;153:189–94. [DOI] [PubMed] [Google Scholar]
  • [8].Yeung HN, Kormos DW. Separation of True Fat and Water Images by Correcting Magnetic Field Inhomogeneity in Situ. Radiology 1986;159:783–6. 10.1017/CBO9781107415324.004. [DOI] [PubMed] [Google Scholar]
  • [9].Glover GH, Schneider E. Three-point dixon technique for true water/fat decomposition with B0 inhomogeneity correction. Magn Reson Med 1991;18:371–83. 10.1002/mrm.1910180211. [DOI] [PubMed] [Google Scholar]
  • [10].Szumowski J, Coshow WR, Li F, Quinn SF. Phase unwrapping in the three-point Dixon method for fat suppression MR imaging. Radiology 1994;192:555–61. 10.1148/radiology.192.2.8029431. [DOI] [PubMed] [Google Scholar]
  • [11].Xiang QS, An L. Water-fat imaging with direct phase encoding. J Magn Reson Imaging 1997;7:1002–15. 10.1002/jmri.1880070612. [DOI] [PubMed] [Google Scholar]
  • [12].Reeder SB, Pineda AR, Wen Z, Shimakawa A, Yu H, Brittain JH, et al. Iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL): Application with fast spin-echo imaging. Magn Reson Med 2005;54:636–44. 10.1002/mrm.20624. [DOI] [PubMed] [Google Scholar]
  • [13].Reeder SB, McKenzie CA, Pineda AR, Yu H, Shimakawa A, Brau AC, et al. Water-fat separation with IDEAL gradient-echo imaging. J Magn Reson Imaging 2007;25:644–52. 10.1002/jmri.20831. [DOI] [PubMed] [Google Scholar]
  • [14].Xiang QS. Two-point water-fat imaging with partially-opposed-phase (POP) acquisition: An asymmetric dixon method. Magn Reson Med 2006;56:572–84. 10.1002/mrm.20984. [DOI] [PubMed] [Google Scholar]
  • [15].Ye Y, Hu J, Haacke EM. Robust selective signal suppression using binomial off-resonant rectangular (BORR) pulses. J Magn Reson Imaging 2014;39:195–202. 10.1002/jmri.24149. [DOI] [PubMed] [Google Scholar]
  • [16].Ye Y, Wu Z, Lewis NA, Fan Q, Haacke EM. Retrobulbar magnetic resonance angiography using binomial off-resonant rectangular (BORR) pulse. Magn Reson Med 2015;74:1050–6. 10.1002/mrm.25498. [DOI] [PubMed] [Google Scholar]
  • [17].Bastiaansen JAM, Stuber M. Flexible water excitation for fat-free MRI at 3T using lipid insensitive binomial off-resonant RF excitation (LIBRE) pulses. Magn Reson Med 2018;79:3007–17. 10.1002/mrm.26965. [DOI] [PubMed] [Google Scholar]
  • [18].Mikami WK, Kazama T, Sato H, Yokota H, Higashide T, Horikoshi T, et al. Fat suppression strategies in MR imaging of breast cancer at 3.0 T: Comparison of the two-point Dixon technique and the frequency selective inversion method. Jpn J Radiol 2013;31:615–22. 10.1007/s11604-013-0230-8. [DOI] [PubMed] [Google Scholar]
  • [19].Dogan BE, Ma J, Hwang K, Liu P, Yang WT. T1-weighted 3D dynamic contrast-enhanced MRI of the breast using a dual-echo dixon technique at 3 T. J Magn Reson Imaging 2011;34:842–51. 10.1002/jmri.22705. [DOI] [PubMed] [Google Scholar]
  • [20].Clauser P, Pinker K, Helbich TH, Kapetas P, Bernathova M, Baltzer PAT. Fat saturation in dynamic breast MRI at 3 Tesla: Is the Dixon technique superior to spectral fat saturation? A visual grading characteristics study. Eur Radiol 2014;24:2213–9. 10.1007/s00330-014-3189-7. [DOI] [PubMed] [Google Scholar]
  • [21].Kalovidouri A, Firmenich N, Delattre BMA, Picarra M, Becker CD, Montet X, et al. Fat suppression techniques for breast MRI: Dixon versus spectral fat saturation for 3D T1-weighted at 3 T. Radiol Medica 2017;122:731–42. 10.1007/s11547-017-0782-2. [DOI] [PubMed] [Google Scholar]
  • [22].Hancu I, Govenkar A, Lenkinski RE, Lee SK. On shimming approaches in 3T breast MRI. Magn Reson Med 2013;69:862–7. 10.1002/mrm.24307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Rahbar H, Partridge SC, DeMartini WB, Gutierrez RL, Parsian S, Lehman CD. Improved B 1 homogeneity of 3 tesla breast MRI using dual-source parallel radiofrequency excitation. J Magn Reson Imaging 2012;35:1222–6. 10.1002/jmri.23571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Winkler SA, Rutt BK. Practical methods for improving B 1 + homogeneity in 3 tesla breast imaging. J Magn Reson Imaging 2015;41:992–9. 10.1002/jmri.24635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Skinner TE, Reiss TO, Luy B, Khaneja N, Glaser SJ. Reducing the duration of broadband excitation pulses using optimal control with limited RF amplitude. J Magn Reson 2004;167:68–74. 10.1016/j.jmr.2003.12.001. [DOI] [PubMed] [Google Scholar]
  • [26].Skinner TE, Reiss TO, Luy B, Khaneja N, Glaser SJ. Application of optimal control theory to the design of broadband excitation pulses for high-resolution NMR. J Magn Reson 2003;163:8–15. 10.1016/S1090-7807(03)00153-8. [DOI] [PubMed] [Google Scholar]
  • [27].Kobzar K, Skinner TE, Khaneja N, Glaser SJ, Luy B. Exploring the limits of broadband excitation and inversion: II. Rf-power optimized pulses. J Magn Reson 2008;194:58–66. 10.1016/j.jmr.2008.05.023. [DOI] [PubMed] [Google Scholar]
  • [28].Matson GB, Young K, Kaiser LG. RF pulses for in vivo spectroscopy at high field designed under conditions of limited power using optimal control. J Magn Reson 2009;199:30–40. 10.1016/j.jmr.2009.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Li W, Xu F, Schär M, Liu J, Shin T, Zhao Y, et al. Whole-brain arteriography and venography: Using improved velocity-selective saturation pulse trains. Magn Reson Med 2018;79:2014–23. 10.1002/mrm.26864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Liu H, Matson GB. Radiofrequency pulse designs for three-dimensional MRI providing uniform tipping in inhomogeneous B 1 fields. Magn Reson Med 2011;66:1254–66. 10.1002/mrm.22913. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].Matson GB. Design strategies for improved velocity-selective pulse sequences. Magn Reson Imaging 2017;44:146–56. 10.1016/j.mri.2017.09.006. [DOI] [PubMed] [Google Scholar]
  • [32].Yokoo T, Shiehmorteza M, Hamilton G, Wolfson T, Schroeder ME, Middleton MS, et al. Estimation of hepatic proton-density fat fraction by using MR imaging at 3.0 T. Radiology 2011;258:749–59. 10.1148/radiol.10100659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].Hamilton G, Yokoo T, Bydder M, Cruite I, Schroeder ME, Sirlin CB, et al. In vivo characterization of the liver fat 1H MR spectrum. NMR Biomed 2011;24:784–90. 10.1002/nbm.1622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Yu H, Shimakawa A, McKenzie CA, Brodsky E, Brittain JH, Reeder SB. Multiecho water-fat separation and simultaneous R*2 estimation with multifrequency fat spectrum modeling. Magn Reson Med 2008;60:1122–34. 10.1002/mrm.21737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Ren J, Dimitrov I, Sherry AD, Malloy CR. Composition of adipose tissue and marrow fat in humans by 1H NMR at 7 Tesla. J Lipid Res 2008;49:2055–62. 10.1194/jlr.D800010-JLR200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Kuroda K, Oshio K, Mulkern R V., Jolesz FA. Optimization of chemical shift selective suppression of fat. Magn Reson Med 1998;40:505–10. 10.1002/mrm.1910400402. [DOI] [PubMed] [Google Scholar]
  • [37].Mao J, Hong Y W.W. B, Bidgood WDJ, J.J. S A. M Fat tissue and fat suppression. Magn Reson Imaging 1993;11:385–93. [DOI] [PubMed] [Google Scholar]
  • [38].Rakow-Penner R, Daniel B, Yu H, Sawyer-Glover A, Glover GH. Relaxation times of breast tissue at 1.5T and 3T measured using IDEAL. J Magn Reson Imaging 2006;23:87–91. 10.1002/jmri.20469. [DOI] [PubMed] [Google Scholar]
  • [39].Matson GB. An integrated program for amplitude-modulated RF pulse generation and re-mapping with shaped gradients. Magn Reson Imaging 1994;12:1205–25. 10.1016/0730-725X(94)90086-7. [DOI] [PubMed] [Google Scholar]
  • [40].Nehrke K, Börnert P. DREAM-a novel approach for robust, ultrafast, multislice B1 mapping. Magn Reson Med 2012;68:1517–26. 10.1002/mrm.24158. [DOI] [PubMed] [Google Scholar]
  • [41].Nehrke K, Sprinkart AM, Börnert P. An in vivo comparison of the DREAM sequence with current RF shim technology. Magn Reson Mater Physics, Biol Med 2015;28:185–94. 10.1007/s10334-014-0454-3. [DOI] [PubMed] [Google Scholar]
  • [42].Miyazaki M, Wheaton A, Kitane S. Enhanced fat suppression technique for breast imaging. J Magn Reson Imaging 2013;38:981–6. 10.1002/jmri.23932. [DOI] [PubMed] [Google Scholar]
  • [43].Abe T. Fast fat suppression RF pulse train with insensitivity to B 1 inhomogeneity for body imaging. Magn Reson Med 2012;67:464–9. 10.1002/mrm.23015. [DOI] [PubMed] [Google Scholar]
  • [44].Hua J, Qin Q, Van Zijl PCM, Pekar JJ, Jones CK. Whole-brain three-dimensional T2-weighted BOLD functional magnetic resonance imaging at 7 Tesla. Magn Reson Med 2014. 10.1002/mrm.25055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [45].Zi R, Zhu D, Qin Q. Quantitative T2 mapping using accelerated 3D stack-of-spiral gradient echo readout. Magn Reson Imaging 2020;73:138–47. 10.1016/j.mri.2020.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [46].Miao X, Wu Y, Liu D, Jiang H, Woods D, Stern MT, et al. Whole-brain functional and diffusion tensor MRI in human participants with metallic orthodontic braces. Radiology 2020;294:149–57. 10.1148/radiol.2019190070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [47].Wu D, Liu D, Hsu YC, Li H, Sun Y, Qin Q, et al. Diffusion-prepared 3D gradient spin-echo sequence for improved oscillating gradient diffusion MRI. Magn Reson Med 2020:78–88. 10.1002/mrm.28401. [DOI] [PubMed] [Google Scholar]
  • [48].Liu D, Xu F, Li W, van Zijl PC, Lin DD, Qin Q. Improved velocity-selective-inversion arterial spin labeling for cerebral blood flow mapping with 3D acquisition. Magn Reson Med 2020;84:2512–22. 10.1002/mrm.28310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [49].Landes V, Javed A, Jao T, Qin Q, Nayak K. Improved velocity-selective labeling pulses for myocardial ASL. Magn Reson Med 2020;84:1909–18. 10.1002/mrm.28253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [50].Zhu D, Li W, Liu D, Liu G, Pei Y, Shin T, et al. Non-contrast-enhanced abdominal MRA at 3 T using velocity-selective pulse trains. Magn Reson Med 2020;84:1173–83. 10.1002/mrm.28187. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1642196-supplement-1.docx (3MB, docx)

RESOURCES