NetForum uses cookies to ensure that we give you the best experience on our website. If you continue to use the site, we'll assume that you are happy to receive these cookies on the NetForum website. Read about our cookies.
NetForum Community
Learn. Share. Optimize.
Log in | Sign up now | Submit content | Contact
Go to similar content

1.5T Brain spectroscopy - Soroka University Medical Center

ExamCard
Rosen, Philip Be'er Sheva, Soroka University Medical Center
Shelef, Ilan, M.D. Be'er Sheva, Soroka University Medical Center

1.5T Brain spectroscopy - Soroka University Medical Center, Be'er Sheva, Israel

Download
Login is required to download ExamCards.

ExamCard purpose

This ExamCard contains a 2D PRESS Turbo Spectroscopic Imaging sequence with SENSE that does not require circular REST slabs. The sequence is configured for Release 2.5 using the 16-channel SENSE NeuroVascular coil. The combined acceleration factor is 12 (SENSE=4 and TSI=3) for a scan time of 68 sec.

Introduction

The importance of nuclear magnetic resonance (NMR) spectroscopy rests on the observation that the resonant frequency of a particular nuclear species or isotope in a molecule is not uniform, but tends to vary slightly according to its position and configuration. For example, the resonance of each hydrogen nucleus (proton) depends on its neighbouring atoms and the distribution of the electrons in the molecular orbital. This change in frequency is known as the chemical shift phenomenon. It is the reason for the difference of about 76 Hz at 1.5T between the proton resonances of the methyl groups of choline and N-acetylaspartate.

 

In order to obtain NMR spectra from in-vivo examinations of the brain, the concentrations of metabolites need to be sufficiently high and their molecular weights need to be sufficiently low. Proteins, glycoproteins and other molecules larger than approximately 10 kDaltons tend to have overlapping peaks and short T2 relaxation times to the extent that the NMR absorption peaks become too broad to be observed. In addition, many low molecular weight compounds in the brain have concentrations that are too low to be detected. This includes many important metabolites, such as the neurotransmitters (except glutamate and gamma-aminobutyric acid) and all the hormone messengers. That is why an MR spectrum of the brain has a typically sparse and simple appearance. The well-separated peaks and finer spectral structures that are characteristic of in-vitro NMR are not observed in vivo because of the comparatively low strength of the main magnetic field and the inherent limitations of field homogeneity with live patients. Fortunately those metabolites that can be detected are important indicators of cellular metabolism, and that is why spectroscopy is so useful in tumor grading.

 

A spectrum from a single voxel in the brain is generally insufficient to obtain a complete diagnosis. It is often necessary to compare spectra in regions immediately surrounding the lesion. Studies showed that a property of high-grade gliomas is that the spectra of the adjacent tissues are characteristic of low-grade lesions, whereas in metastases the spectra will appear normal (Delorme and Weber, 2006).

 

It is possible to acquire spectra simultaneously from a two- or three-dimensional array of voxels using a technique similar to magnetic resonance imaging (MRI) known as chemical shift imaging (CSI). An important difference between MRI and CSI however, is that the resolution is obtained only by the application of phase encoding gradients. Frequency encoding is not used because the frequency differences that are detected must only originate from the spectral chemical shift.

 

The main impediment to using CSI acquisitions is that they can take several minutes to acquire because only phase encoding gradients are used. Two techniques are available on the Philips scanner to reduce the scan time to make CSI acquistions feasible. One is Turbo Spectroscopic Imaging (TSI) that shortens the scan time by the size of the turbo factor as in TSE imaging. The other is SENSE and can be used for both CSI and TSI acquisitions.

 

In the description that follows only TSI scans are used, but the planning, acquistion and post-processing for CSI is identical. There are two techniques available for defining the slice in 2D-TSI: Spin-echo TSI (SE-TSI) and Point Resolved Spectroscopy TSI (PRESS-TSI).

Turbo Spectroscopic Imaging

SE-TSI uses a slice-selective 90° excitation pulse followed by a 180° refocusing pulse with one selection gradient to define the slice. Now for the purposes of this discussion a 2D slice should be thought of as a volume just as a slice of bread occupies a volume. Fig. 1 shows the scan planning for a 2D SE-TSI sequence. The acquired volume here is the defined FOV of the slice (delineated by the edge of the voxel grid) multiplied by its thickness. The shimmed volume from which the spectra are obtained is the volume of acquisition (VOI) shown in Fig. 1 as the yellow rectangle within the FOV. The SE-TSI sequence therefore needs to use circular REST slabs (shown as blue bands around the head) to suppress the signal from outside the VOI, in particular to suppress the lipid signal from the skull. This inevitably reduces further the useful area of the VOI that is already constrained because of geometric restrictions between the rectangular VOI and the oval shape of the head.

 

PRESS-TSI uses a slice-selective 90° excitation pulse followed by two 180° refocusing pulses and orthogonal selection gradients to define the acquisition volume. In PRESS-TSI the shimming and acquisition is now defined by the VOI. The advantage of PRESS-TSI is that the VOI is so well-defined that there is now no need for the REST slabs. This makes the planning much easier as shown in Fig. 2 in which the VOI is outlined in yellow. The limit of the FOV is the edge of the voxel grid as in Fig.1.

Fig. 1: SE-TSI planning The appearance of circular REST slabs superimposed on the FOV and VOI (in yellow) in SE-TSI planning.Fig. 2: PRESS-TSI planning The display of the FOV and VOI (in yellow) in PRESS-TSI planning.
Fig. 1: SE-TSI planning
Fig. 2: PRESS-TSI planning
The appearance of circular REST slabs superimposed on the FOV and VOI (in yellow) in SE-TSI planning.
The display of the FOV and VOI (in yellow) in PRESS-TSI planning.

ExamCard overview

sTSI-PRESS for NV-16
Scan 1SURVEY/FFE2
Scan 2Ref_NV_16
Scan 3sTSI_PRESS

The sTSI_PRESS sequence was built from the Philips standard sTSI_SE sequence by making the following changes:

  • Scan Mode from MS to 2D.
  • VOI Selection from Slice to Volume.
  • Method from SE to PRESS.
  • RF pulse from Sharp to Normal.
  • The SENSE factor was set to 2 in the P and M directions.
  • The voxel size was set to 10 × 10 × 15mm (P×M×S).
  • The TR was set to 1600 msec.

 

The turbo factor was left unchanged at 3, and the spectral resolution at 4Hz. The combination of TSI and SENSE resulted in a combined acceleration factor of 12 (SENSE=4 x TSI=3). (For other coils such as the SENSE-6 Head coil, the combined SENSE factor should be reduced to 1.8x1.8 in the P and M directions).

 

It is recommended to transfer the spectroscopic sequence into the current ExamCard only after all the axial scans have been completed. This is done to avoid changing the Foot-Head (FH) coordinate of these imaging sequences. Then proceed as follows:

 

  1. Change the scan geometry to that used for the axial scans. This ensures that the VOI has the same angulation as the imaging sequence.
  2. Plan the sTSI_PRESS sequence on a the selected axial image using the leftmost window of the three localiser subscreens. The position on the FH axis is changed using a sagittal or coronal image in either of the two remaining subscreens.
  3. Adjust the size of the VOI according to the dimensions of the brain and the lesion to be examined.

 

The change from SE to PRESS alters the direction of the chemical shift artifact from the slice to the in-plane direction. The use of PRESS also requires the substitution of the Normal for the Sharp RF pulse which reduces the bandwidth. This can be seen in Fig.3 which is an enlarged view of the TSI scan planning. The white rectangle marks the extent of the NAA-Choline chemical shift. The shift is quite small relative to the voxel size and can be ignored at 1.5T.

Fig. 3: TSI-PRESS planning The extent of the NAA-Cho chemical shift artifact in the sTSI_PRESS sequence.
Fig. 3: TSI-PRESS planning
The extent of the NAA-Cho chemical shift artifact in the sTSI_PRESS sequence.

 

In Fig 4, we can see the VOI and the acquistion voxels from a SENSE TSI scan defined as they would appear in the Spectroscopy package. The voxels outlined in blue were selected for post-processing. The results of the four processed voxels (shaded in yellow) are compared in Fig. 5 and Fig. 6 for a SE-TSI and a PRESS-TSI acquisition, respectively. Both of the sequences had the same SENSE (=2×2) and TSI (=3) factors.

Fig. 4: TSI voxels for post-processing The voxels outlined in blue were processed, and those shaded in yellow are displayed in Fig. 5 and Fig. 6 for SE-TSI and PRESS-TSI scans, respectively.<br>
Fig. 4: TSI voxels for post-processing
The voxels outlined in blue were processed, and those shaded in yellow are displayed in Fig. 5 and Fig. 6 for SE-TSI and PRESS-TSI scans, respectively.
Fig. 5 SE-TSI SENSE Spectra The four SE-TSI spectra from the voxels shaded in yellow in Fig. 4 displayed with the same relative positions.<br>
Fig. 5 SE-TSI SENSE Spectra
The four SE-TSI spectra from the voxels shaded in yellow in Fig. 4 displayed with the same relative positions.
Fig. 6 PRESS-TSI SENSE Spectra The four PRESS-TSI spectra from the voxels shaded in yellow in Fig. 4 displayed with the same relative positions.
Fig. 6 PRESS-TSI SENSE Spectra
The four PRESS-TSI spectra from the voxels shaded in yellow in Fig. 4 displayed with the same relative positions.

 

It can be seen that when PRESS-TSI and SE-TSI are combined with SENSE, the PRESS-TSI sequence is less sensitive to artifacts than the SE-TSI. In particular, the region downfield from the NAA peak in the region of 1.6 ppm where spurious absorptions are sometimes seen.

 

Below follow two clinical examples in which SENSE PRESS-TSI was used.

Spectra of a non-tumorous lesion

The first case is a patient suffering from encephalitis. The scan planning is shown in Fig. 7 as it appears in the Spectroscopy Post-Processing package. It shows how close the VOI (shown in green) can be placed to the skull to include as much affected tissue as possible. This is feasible because the PRESS-TSI has such a well-defined profile. It does not require the use of circular REST slabs to suppress unwanted signals. The resulting spectra are shown in Fig. 8. Notice that the creatine and choline peaks are well-separated indicating good shimming, and that there is little sign of contaminating lipid peaks beyond the level of the background noise.
Fig. 7: Post-processing planning The voxels outlined in blue were processed, and the four voxels shaded in yellow are displayed in Fig. 8 below.<br>
Fig. 7: Post-processing planning
The voxels outlined in blue were processed, and the four voxels shaded in yellow are displayed in Fig. 8 below.
Fig. 8: SENSE PRESS-TSI Spectra Spectra of the processed voxels shaded in yellow from Fig. 7. The relative positions of the four shaded voxels in Fig 7 are maintained in the display here.
Fig. 8: SENSE PRESS-TSI Spectra
Spectra of the processed voxels shaded in yellow from Fig. 7. The relative positions of the four shaded voxels in Fig 7 are maintained in the display here.

 

The voxels over the lesion still show NAA and choline peaks although reduced in intensity. An additional feature of TSI and CSI is that maps can be generated for each visible resonance in which each voxel is given a colour that is proportional to the concentration of a specific metabolite peak in that voxel. Increasing concentration is represented by a transition from red to yellow. This map is then overlaid on a standard image. This is shown in Fig. 9 for the NAA peak and in Fig 10 for the Choline/NAA peak ratio.

Fig. 9: NAA map Metabolite map showing the variation in NAA concentration in the VOI defined in Fig. 7Fig. 10: Cho/NAA ratio map Metabolite map showing the ratio of the Cho to NAA concentrations in the VOI defined in Fig. 7
Fig. 9: NAA map
Fig. 10: Cho/NAA ratio map
Metabolite map showing the variation in NAA concentration in the VOI defined in Fig. 7
Metabolite map showing the ratio of the Cho to NAA concentrations in the VOI defined in Fig. 7

 

The metabolic maps clearly show the regional variations in NAA and choline concentrations in the VOI defined for the post-processing as shown in Fig 7. It can be seen in Figs. 8, 9 and 10 that there is a distinct change in the distribution of the NAA and choline intensities. The persistence of the NAA peak and the relatively low intensity of the choline peak in the lesion are not associated with a tumor. The diagnosis of encephalitis was made based on clinical symptoms and laboratory findings (herpes simplex virus).

Spectra of a tumorous lesion

The post-processing delineations of the VOI and the displayed voxels are shown in Fig. 11. As before, the VOI (shown in green) could be placed next to the skull without contamination from lipid signals. The resulting spectra from the six voxels shaded in yellow are shown in Fig. 12.
Fig. 11: Post-processing planning The voxels outlined in blue were processed. The six voxels shaded in yellow are displayed in Fig. 12 below.<br>
Fig. 11: Post-processing planning
The voxels outlined in blue were processed. The six voxels shaded in yellow are displayed in Fig. 12 below.
Fig. 12: PRESS-TSI SENSE Spectra Spectra of the processed voxels shaded in yellow from Fig. 11. The relative positions of the six shaded voxels in Fig. 11 are maintained in the display here.
Fig. 12: PRESS-TSI SENSE Spectra
Spectra of the processed voxels shaded in yellow from Fig. 11. The relative positions of the six shaded voxels in Fig. 11 are maintained in the display here.

 

The voxels close to the lesion show high choline and low NAA peaks associated with a primary neoplastic lesion. The NAA and NAA/Choline maps shown in Fig. 13 and Fig. 14 illustrate this distribution. The tumor was later confirmed as a lymphoma by stereotactic biopsy.

Fig. 13: NAA Map Metabolite map showing the variation in NAA concentration over the processed VOI defined in Fig. 11.Fig. 14: Cho/NAA ratio map Metabolite map showing the ratio of the Cho to NAA concentrations over the processed VOI defined in Fig. 11.
Fig. 13: NAA Map
Fig. 14: Cho/NAA ratio map
Metabolite map showing the variation in NAA concentration over the processed VOI defined in Fig. 11.
Metabolite map showing the ratio of the Cho to NAA concentrations over the processed VOI defined in Fig. 11.

Summary

 

The SENSE PRESS-TSI sequence generates excellent spectra that are generally free from artifacts without the need for circular REST slabs. The acquisition time is only 68 sec. for a voxel resolution of 10×10×15 mm. Furthermore, it is our experience that the circular REST slabs can be omitted in all CSI and TSI scanning when PRESS is used because of its excellent volume definition.

Reference

Delorme S and Weber MA

Applications of MRS in the evaluation of focal malignant brain lesions.

Cancer Imaging (2006) 6: 95-99.



This content has been made possible by NetForum Community.
Share this on: Share your link in twitter Share your link in facebook Share your link on LinkedIn Print Rate this article: Log in to vote

 
Rating:
Votes:
1
Views:
3918
Added:
Jul 2, 2009

Rate this:
Log in to vote
 

ExamCard
Achieva 1.5T
Release 2
Brain, CSI, Neuro, PRESS, Spectroscopy
 

Clinical News
Best Practices
Case Studies
Publications and Abstracts
White Papers
Web seminars and Presentations
ExamCards
Protocols
Application Tips and FAQ
Training
Try an Application
Business News
Case Studies
White Papers
Web Seminars and Presentations
Utilization Services
Contributing Professionals
Contributing Institutions
Become a Contributor