Improved whole brain water suppression efficiency with four-pulse WET in echo-planar spectroscopic imaging (EPSI) at 7 tesla
Graeme A. Keith1 and David A. Porter1

1Imaging Centre of Excellence, University of Glasgow, Glasgow, United Kingdom


Water suppression in MR spectroscopic imaging can be sensitive to variations in B1 in the sample, such as are present at 7 tesla. This work compares two versions of the WET water suppression method, the standard 3-pulse method and the extended 4-pulse method which is expected to be less sensitive to B1 variation. It is found that the 4-pulse method provides a greater consistency of water suppression efficiency across a range of B1 in both phantoms and the brain at 7T.


Suppression of the water signal in 1H MRS is necessary due to lower concentration of tissue metabolites than of water. The WET technique1 is an established method in MRS/MRSI for suppression of the dominant water signal. The method was developed to be insensitive to a range of B1 and T1 values, but with the greater B1-inhomogeneity at higher field-strength the efficiency of the technique over a larger field-of-view with multi-voxel MRSI methods decreases. In this work, to overcome the limitations of the existing implementation at 7T, the standard method was compared with the extended, four-pulse WET method for water suppression efficiency in a phantom and in the brain.


Two versions of WET were implemented in a spin-echo EPSI sequence2. These were the standard WET, consisting of three Gaussian pulses of nominal flip-angle 89.2°, 83.4° and 160.8° and a four-pulse version with nominal flip-angles 81.4°, 101.4°, 69.3° and 161.0° as optimised in Ogg1 to reduce the residual water magnetisation in the presence of greater B1inhomogeneity. Data for comparison were acquired on a Terra 7T scanner (Siemens Healthcare, Erlangen) with a single-transmit, 32-channel receive head coil (Nova medical, Wilmington, MA). A range of values for the water suppression bandwidth were used, from 35 to 140Hz, with 5Hz steps. Sequence parameters were TR/TE = 2000/13.2ms, FoV = 250mm, matrix size 128x112, 128 echoes with echo spacing of 620μs corresponding to a spectral bandwidth of 1.6kHz. Scans were run for both WET methods and with no suppression. For phantom experiments, the scanner manufacturer's standard spectroscopy phantom was used, which contains a solution of sodium acetate (8.2g/1000g H2O) and a lactic-acid lithium salt (9.6g/1000g H2O). Figure 1(a) shows a flip-angle map, from a dual-angle B1-mapping sequence, in which the poor B1 profile across the phantom is shown. The range of flip-angles achieved was 46.3° to 124.4°, where the nominal flip-angle was 90°. The results were assessed as water suppression efficiency, WSeff, expressed as percentage, and described by:

$$WS_{eff}= \frac{I_{un}-I_{su}}{I_{un}}*100$$

where Iun is the intensity of the unsuppressed water peak and Isu the intensity of the suppressed water peak. The experiment was repeated in vivo using the same RF coil, for both WET methods and pulse bandwidths of 35, 65 and 135 Hz. Sequence parameters were TR/TE = 2000/12.2ms, FoV = 480mm, matrix size 64x64, 256 echoes with spacing 340μs corresponding to spectral bandwidth of 2.9kHz. Figure 1(b) shows the in vivo flip-angle map, with a range of flip-angles from 42.4° to 133.2°. Results were again expressed as water suppression efficiency.


Figure 2(a) and (b) show water suppression efficiency colour maps for both the three-pulse and four-pulse WET method, for the above stated range of bandwidths. Figure 3(a) and (b) show the colour maps for the in vivo application of the two WET methods, for the bandwidths 35, 65 and 135Hz. The colour scale used is matched between all maps and the in vivo data were interpolated. Figure 4 shows suppressed water peaks at 3 points in the phantom with both methods at 135Hz.


The efficiency maps for the phantom in Figure 2 show a clear spatial difference in the effectiveness of the two methods. Though the standard 3-pulse method performs well in some areas of the phantom, the efficiency drops in areas of more extreme flip-angle. The maps for the 4-pulse method show a more even distribution of high suppression efficiency across a range of flip-angles. The in vivo efficiency maps in Figure 3 show that both methods fail with the use of a low bandwidth, 35Hz, with voxels in the central area around the ventricles and lateral areas of the brain with more extreme values of flip angle poorly suppressed. With the higher bandwidths however, the colour maps for the 3-pulse method continue to show some areas of lower suppression efficiency, whereas the 4-pulse method performs well, with the exception of the centre of the brain, where flip-angles are in excess of 120°. In routine 7T scanning, it is expected that the transmit voltage may be adjusted to avoid such excessively high flip-angles, which was not done here. With this scaling approach, it is expected that the 4-pulse WET will provide a sufficient spatial consistency of water suppression levels, despite the variation of B1 at 7T with a single-transmit coil.


This work has compared the effectiveness of 3-pulse and 4-pulse versions of WET water suppression in the presence of high B1-inhomogeneity at 7T. The 4-pulse method, optimised for insensitivity to B1 variation shows greater consistency of water suppression across a range of flip-angles in both phantoms and in vivo.


No acknowledgement found.


1. Ogg, R.J., P.B. Kingsley, and J.S. Taylor, WET, a T1- and B1-insensitive water-suppression method for in vivo localized 1H NMR spectroscopy. J Magn Reson B, 1994. 104(1): p. 1-10.

2. Mansfield, P., Spatial mapping of the chemical shift in NMR. Magn Reson Med, 1984. 1(3): p. 370-86.


Figure 1: Flip-angle maps acquired on (a) a Siemens’ spectroscopy phantom with a nominal 90° flip angle. The measured flip-angles range from 46.3° to 124.4°, and (b) in vivo, where measured flip-angles range from 42.4° to 133.2° across the whole brain.

Figure 2: Water suppression efficiency maps in the scanner manufacturer's standard spectroscopy phantom for the range of bandwidths 35Hz to 140Hz with 5Hz steps using, (a) 3 pulse WET, and (b) 4 pulse WET.

Figure 3: In vivo water suppression efficiency maps in the brain for (a) 3 pulse WET, and (b) 4 pulse WET for pulse bandwidths of 35, 65 and 135 Hz.

Figure 4: (a-c) Spectra of the suppressed water peak with both the 3-pulse and 4-pulse WET method at 3 points in the phantom, and (d) the spatial location of each of the 3 points a, b and c on the first echo from the non-water suppressed dataset. The points were chosen to correspond to values on the flip-angle map of approximately 60°, 90° and 120°.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)