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

Attention valued NetForum members:

Due to evolving technology needs and global privacy regulations, we have made the hard decision to suspend the NetForum User Community platform on Monday, December 9, 2019.

After this date, the current NetForum can no longer be reached. Please click here for more information about this decision, what happens next and how to stay in touch with us about the future of the community.

Go to similar content

10 tips for prostate spectroscopy

Application Tip
de Kok, Wendy Philips Healthcare

Introduction

Over the last few years, a growing interest in prostate spectroscopy is seen. It may be a helpful tool for diagnostics as it provides non-invasive monitoring of metabolite changes, that might precede anatomical changes in pathological processes.

 

This document provides tips for users to perform prostate spectroscopy.

 

General information on how to perform spectroscopy can be found in the Application Tip:  "10 tips for spectroscopy":

Tip 1: Patient preparation - measures to reduce random motion

The position of the prostate is subject to motion, caused by bowel motion and breathing and by the amount of bladder filling. This motion can cause the prostate to move away from the spectroscopy scan volume.

 

To avoid motion induced partial volume effects:

  • tell patient to empty the bladder before the examination.
  • avoid scanning directly after lunch.

 

Additional measures to reduce bowel motion are:

  • use intra-muscular glucagon.
  • cleanse rectum 30 minutes before examination.

 

If a spectrum is not good, and motion of the prostate is the most likely cause, a fast imaging scan can be performed. These images can be used in planscan to check the position of the PRESS-volume. The spectroscopy shim settings should be used for the imaging scan, and can be re-used for later spectroscopy scans.

Position of PRESS at t=0 Position of PRESS at t=2:30 min Prostate position is changed because of bowel motion. It moved out of the PRESS-volume.Bladder motion Left planscan image: acquired before spectro scan.
Right planscan image: acquired after spectro scan.
Bladder filling has pushed the prostate out of the PRESS-volume.
Position of PRESS at t=0
Position of PRESS at t=2:30 min
Bladder motion
Prostate position is changed because of bowel motion. It moved out of the PRESS-volume.
Left planscan image: acquired before spectro scan. Right planscan image: acquired after spectro scan. Bladder filling has pushed the prostate out of the PRESS-volume.

Tip 2: Coil selection

Any coil can be used for spectroscopy. Coil element selection is limited in non-Achieva systems, the maximum number of elements is 1. This limitation is not valid for Achieva systems, where all coil elements of a SENSE or Synergy coil can be selected.

 

If multi-element acquisition is performed, it is important to select only those elements that are close to the prostate, as other elements will add only noise to the signal.

 

Coil selection will be dependent on the imaging coil used, as it is most convenient to use the same coil for both imaging and spectroscopy.

 

Advantages and disadvantages per coil:

 

  • SENSE Flex-M or L:

Relatively easy positioning, and intermediate signal-to-noise ratio for imaging. Penetration of the Flex-M coil might not be sufficient for spectroscopy, dependent on patient size. For multi-element spectroscopy acquisition, both elements are in optimal position.

 

  • Endo-rectal coil:

Very good signal-to-noise ratio for both imaging and spectroscopy, but its use can cause patient discomfort and deformation of the prostate gland due to pressure of the coil.

Note that the air-inflated balloon of the Endo-rectal coil causes susceptibility and might give problems in spectroscopy.

 

  • SENSE Body coil:

Very good signal-to-noise ratio for imaging, but location of a single element with respect to the prostate is not optimal for spectroscopy. Selection of all elements for spectroscopy is not optimal, as it will increase noise level.

Influence of element selection Spectrum above acquired with element 1 and 4 of SCC,
spectrum below acquired with element 2 and 4 of SCC.
Increased SNR, because element 2 was positioned correctly.2DSI Acquired with Endo coil. High signal-to-noise ratio.
Influence of element selection
2DSI
Spectrum above acquired with element 1 and 4 of SCC, spectrum below acquired with element 2 and 4 of SCC. Increased SNR, because element 2 was positioned correctly.
Acquired with Endo coil. High signal-to-noise ratio.

Tip 3: Patient positioning - feet first or manual mode

The maximum allowed table stroke for spectroscopy is smaller than in imaging. With a head-first positioning of the patient, this might lead to an impossible tabletop position for spectroscopy, if table movement inwards is required between the anatomical scans and the spectroscopy acquisition.

 

The impossible table topposition is usually avoided by positioning the patient feet-first. Position the patient such that the anatomy of interest remains within the marker that indicates the maximum imaging table stroke.

(Please note that a tall patient's feet might stick out over the edge of the table.)

 

Feet-first postioning is NOT allowed in combination with the Endo-rectal coil.

 

The impossible tabletop position can be ignored:

  • plan one imaging slice in the FH-offcenter that is required for the spectroscopy acquisition.
  • start imaging scan.
  • travel to scanplane.
  • stop imaging scan.
  • switch to manual mode on the patient interface control unit.
Patient interface control unit Switch to manual mode, next to the emergency stop button.
Patient interface control unit
Switch to manual mode, next to the emergency stop button.

Tip 4: Planscan

To practice prostate spectroscopy, always start with single voxel acquisitions to train how to perform good planning of the VOI. If the resulting spectrum is fine, 2DSI can be performed.

 

Always acquire or load anatomical images in at least two orthogonal directions to check the position and size of the three-dimensional VOI:

  • The VOI-size should be as large as possible for optimal signal-to-noise ratio.
  • Make sure that only prostate tissue is included.
  • Make sure that the VOI doesn't include air-tissue interfaces.
  • Keep the VOI away from the large rim of fat on the frontal side of the prostate gland.

 

Note that the planscan-VOI represents the VOI of signals at +/- 2 ppm (this setting is inherited from brain-spectroscopy, where 2.02 ppm represents the NAA-volume).

The signals from other metabolites come from slightly shifted volumes. The fat volume is shifted.

 

The VOI feet-head size determines the slice thickness in spectroscopic imaging. The size of the VOI should be planned as careful as described before. Additionally, the spectroscopic imaging field of view should be planned:

  • Make sure that the CSI-slice corresponds to an anatomical imaging slice. This image is used as anatomical underlay in post-processing.
  • Always use full volume selection as the anatomy extends beyond the SI-FOV. Use of slice selection will result in fold-over. 
  • Don't use circular REST-slabs for outer volume suppression as signal from the small PRESS-volume is suppressed.

 

Turbo Spectroscopic imaging is not used in the prostate. The J-coupling evolution of citrate is very complex and it is impossible to generate multiple echoes in which the citrate signal has equal phase.

Planning single voxel Planning 2DSI The PRESS-volume is displayed, along with the SI-FOV that is spatially encoded.
Planning single voxel
Planning 2DSI
The PRESS-volume is displayed, along with the SI-FOV that is spatially encoded.

Tip 5: Shimming

Shimming of the volume of interest is performed. The quality of shimming is largely dependent on the planning of the volume (see tip 4).

 

Secondly, the selected method of shimming is important:

  • Automatic shimming is best performed. It is most accurate in a small VOI, especially in those regions of the body where random motion occurs. It requires a few minutes.
  • HOS-shimming only requires 30 seconds, but shimming will be less optimal due to limited phase information that is collected.
Autoshim vs HOS Upper spectrum acquired with autoshim. It has more narrow linewidth than the lower spectrum, that was acquired with HOS (first order). Linewidth of Cho-resonance is doubled.
Autoshim vs HOS
Upper spectrum acquired with autoshim. It has more narrow linewidth than the lower spectrum, that was acquired with HOS (first order). Linewidth of Cho-resonance is doubled.

Tip 6a: Fat suppression

The signals of water and fat are several orders of magnitude larger than the signals of the metabolites, and should be suppressed.

 

Fat suppression is performed by spectral selective inversion. The suppression frequency is set to the fat precession frequency. The suppression window determines the bandwidth of frequencies that is inverted by the inversion pulse. It:

  • should be wide enough to suppress the entire fat peak.
  • should not suppress the peak of citrate at 2.6 ppm (it should not suppress signals at ppm > 2.2 - 2.3).
  • is dependent on the shim quality: larger linewidth requires larger suppression window.

If the required suppression window is large, the suppression center frequency can be shifted to lower frequencies to avoid suppression of the citrate peak.

 

Fat suppression optimization is performed during the preparation phases. Several fat spectra are acquired, with decreasing inversion delay time. Start inversion delay time is set in the preset, and should be larger than the expected optimum (250 - 300 ms).

The spectrum with lowest residual signal should be selected and the corresponding inversion delay time is used in the actual spectroscopy scan.

Note that the resulting spectra might show water signal as well, as the water signal is very strong. It is important to zoom out to detect the residual fat signal.

 

If the TR is long enough for full T1-relaxation of fat, the optimal inversion delay time can also be estimated from the T1-relaxation constant of fat. Enter the desired value as start value in the preset: Optimization can be skipped.

 

Note that fat suppression is not always required if fat tissue is not included in the PRESS-volume. This occurs only if the PRESS-volume is very small.

Fat suppression optimization Upper window: real spectra showing water signal and residual fat.
Lower window: zoomed modulus spectra showing that spectrum 3 has least residual fat (see markers).
Fat suppression optimization
Upper window: real spectra showing water signal and residual fat. Lower window: zoomed modulus spectra showing that spectrum 3 has least residual fat (see markers).

Tip 6b: Water suppression

Water suppression is most easily performed with a basing pulse. Additional gradients and inversion pulses are added to the volume selection sequence, to suppress the signal of water. Use of the basing pulse is only possible in scans with long TE (>  +/- 100ms).

 

The suppression frequency is set to the water frequency, the suppression window determines the bandwidth of the inversion pulse and thus the bandwidth of frequencies that are suppressed.

The gradient strength and duration can be chosen. Higher strength and/or longer duration lead to more complete water suppression, but also require more time, and thus longer TE.

 

Additional water suppression can be performed by using spectral selective excitation or inversion, but both methods require optimization. Combination of two optimization steps, for both fat and water suppression, requires some skills in setting up the protocol parameters:

  • timing of the various inversion and excitation pulses can be critical. In general, shorter time is required if suppression window is enlarged.
  • Maximum allowed water suppression window is quite large, as the closest metabolite of interest is located at 3.2 ppm. Signals at ppm>3.6-3.7 can be suppressed.
  • The suppression frequency can be shifted if a larger window is required.

 

During optimization of the suppression angle it can be seen that the main portion of the water signal is suppressed by the excitation prepulse and that the remainder of the water signal is suppressed by the basing pulse.

 

Suppression parameters If water and fat suppression are combined, the required suppression windows are usually large. The suppression frequency is shifted to avoid suppression of the metabolites.Water suppression optimization Both time-domain signals and resulting spectra. Combination of exc. and basing leads to good water suppression in spectrum 4 and higher.
Suppression parameters
Water suppression optimization
If water and fat suppression are combined, the required suppression windows are usually large. The suppression frequency is shifted to avoid suppression of the metabolites.
Both time-domain signals and resulting spectra. Combination of exc. and basing leads to good water suppression in spectrum 4 and higher.

Tip 7: Repetition time and echo time

The repetition time should be long enough to allow good T1-relaxation of the metabolite signals for optimal signal-to-noise ratio. T1-relaxation constants of the metabolites are long, and very long TR would be needed for complete relaxation. This will lead to unacceptable long scan times.

Therefore, the TR should be chosen such that a good compromise between scan time and signal-to-noise ratio is found: 1200 - 1500 ms.

The last bit of T1-relaxation is slow, and gain in signal by increasing the TR is not very noticeable anymore. Signal is better increased by increasing the number of measurements.

 

The echo time must be chosen such that citrate signal is best detected. Citrate is a strongly coupled system, and its signal splits into a quartet. The complex J-coupling evolves such that citrate is best detected at TE = 120 - 130 ms at 1.5T and TE = 100 - 107 ms at 3.0T (citrate signal has negative sign at 100 - 107 ms).

Effect of TR on SNR Upper spectrum: TR 1500 ms, relative SNR citrate 30. Lower spectrum: TR 1200 ms, relative SNR citrate 24. Signal increases almost linearly with TR, since TR is relatively short in relation to T1-relaxation time.3.0T evolution of citrate signal Three spectra, acquired with TE 100, 107 and 110 ms. 
Complex coupling is clearly seen. The citrate signal is best detected at 100-107 ms.3.0T evolution of citrate signal Three spectra, acquired with TE 120, 140 and 263 ms. 
Complex coupling is clearly seen. The citrate signal is also well defined at 263 ms, but SNR will be lower due to T2-decay.
Effect of TR on SNR
3.0T evolution of citrate signal
3.0T evolution of citrate signal
Upper spectrum: TR 1500 ms, relative SNR citrate 30. Lower spectrum: TR 1200 ms, relative SNR citrate 24. Signal increases almost linearly with TR, since TR is relatively short in relation to T1-relaxation time.
Three spectra, acquired with TE 100, 107 and 110 ms. Complex coupling is clearly seen. The citrate signal is best detected at 100-107 ms.
Three spectra, acquired with TE 120, 140 and 263 ms. Complex coupling is clearly seen. The citrate signal is also well defined at 263 ms, but SNR will be lower due to T2-decay.

Tip 8: Number of measurements

The number of measurements that is required for single voxel depends on the PRESS volume size. In general, a high number of measurements is required to obtain sufficient signal-to-noise ratio.

In comparison: the number of measurements for an equal voxel size in brain spectroscopy is lower than for prostate spectroscopy, as the concentration of the metabolites in the prostate is lower than the concentration of the metabolites in the brain.

 

A good default for a volume of 20 x 30 x 25 mm is 256 NSA.

 

For spectroscopic imaging the total number of measurements is also determined by the matrix size as every phase encoding step acquires signal from the entire PRESS-volume. Usually the PRESS volume is that small that multiple NSA are still required. Some guidelines:

  • Don't use very low scan matrix as this will reduce voxel definition, and:
  • Low scan matrix = low nr of phase encodings = low SNR. A good average scan matrix is 12x12 or 16x16 (also FOV-dependent), and:
  • Depending on coil the minimal nr of NSA should be 3 (Endo-coil) or 4 (other coils).

 

A general rule of thumb for signal-to-noise ratio in prostate MRS (1.5T):

If the voxel size is smaller than 1 cc, the required scan time is +/- 20 minutes.

 

Effects of NSA on SNR SVS, TE 130 ms. Black spectrum: 256 NSA, scan time 5:31 min. Red spectrum: 384 NSA, scan time 8:16 min. SNR increases with increasing nr. of measurements.
Effects of NSA on SNR
SVS, TE 130 ms. Black spectrum: 256 NSA, scan time 5:31 min. Red spectrum: 384 NSA, scan time 8:16 min. SNR increases with increasing nr. of measurements.

Tip 9: Frequency lock

Spectroscopy is a technique that reveals the frequencies of nuclei in different metabolites. It is therefore important that f0, detected in the preparation phase, doesn't drift during a spectroscopy scan, as this will result in linebroadening of the peaks.

 

A frequency lock can be selected, that performs a new f0-determination in every TR. This technique is very robust in non-moving tissue. But random motion (see tip 1) is present in prostate spectroscopy and frequency lock should not be used. The detected linewidth will in general be worse, and in some cases a spectrum will not be found at all.

Frequency lock comparison Upper spectrum acquired without frequency lock. Lower spectrum acquired with frequency lock. Broader lines are seen in the lower spectrum, but spectral quality is acceptable.Frequency lock comparison Upper spectrum acquired without frequency lock. Lower spectrum acquired with frequency lock. f0-determination failed due to motion, and the resulting spectrum cannot be used.
Frequency lock comparison
Frequency lock comparison
Upper spectrum acquired without frequency lock. Lower spectrum acquired with frequency lock. Broader lines are seen in the lower spectrum, but spectral quality is acceptable.
Upper spectrum acquired without frequency lock. Lower spectrum acquired with frequency lock. f0-determination failed due to motion, and the resulting spectrum cannot be used.

Tip 10: Post-processing in SpectroView

Both SVS and 2DSI can be processed in SpectroView. Fitted spectra, tables of results, metabolite maps and ratio maps in color overlay can be generated. The quality of the results depends on the quality of the original spectrum and the processing steps that are executed.

 

Default processing scripts are provided and will lead to good results in most cases. If quality of the processed spectra is not satisfying, the pre-defined scripts can easily be changed to improve the results of the fitting routine.

 

Several effects can affect the quality of the processed spectra:

  • Large residual fat signal can reduce the quality of the fit. The analysis range can be chosen such that fat is excluded.
  • If residual fat is not present, increasing the analysis range can improve the quality of the baseline correction.
  • Broader linewidth can cause the fit to fail if too many baseline polynomials are used. Improve by reducing the number of polynomials. Always check the quality of baseline correction in the spectral display.
  • Lock relative frequency and/or lock relative width can be used if one or more peaks cannot be detected properly due to low signal.
  • Lock relative frequency and/or lock relative width should not be used if fine control of the fitting is required.

 

 

Default script  Upper spectrum: Increased analysis range. Lwer spectrum: lock relative frequency is not used. The fit for choline and creatine is more accurate.
Default script
Upper spectrum: Increased analysis range. Lwer spectrum: lock relative frequency is not used. The fit for choline and creatine is more accurate.


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:
4433
Added:
Jul 18, 2005

Rate this:
Log in to vote
 

Application Tip
Achieva 1.5T, Achieva 3.0T, Ingenia 1.5T, Ingenia 3.0T, Intera 1.5T
Release 1, Release 10, Release 11, Release 12, Release 2, Release 3, Release 4, Release 5, Release 9
Body, Prostate, 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