Abstract Archives of the RSNA, 2010
Namkug Kim PhD, Presenter: Nothing to Disclose
Young-Hoon Sung, Abstract Co-Author: Nothing to Disclose
Andrew Prescot, Abstract Co-Author: Nothing to Disclose
In Kyoon Lyoo, Abstract Co-Author: Nothing to Disclose
Perry F. Renshaw, Abstract Co-Author: Nothing to Disclose
The outcome of this study suggests that optimal weights and phase alignment on individual spectra enhances the quality of MR MEGAPRESS spectroscopy.
It is not easy to get a reliable gamma aminobutyric acid (GABA) measurement for its low signal-to-noise ratio (SNR) in cortical grey matter (~1mM in vivo) and the superimposition with signals of other metabolites. J-difference editing based upon Point Resolved Spectroscopy with MEGA suppression (MEGAPRESS) has been commonly used. To increase SNR of GABA signal of MEGAPRESS, the optimized weights combination of the data from each phased array coil and phase alignment on individual spectra were performed to increase SNR and compensate motion artifact and magnet instability.
Single voxel spectroscopy data was acquired from a 3T MRI (Siemens MAGNETOM Trio) with 12 channel head coil. 8 scans on human brain were performed by MEGAPRESS with TR 2s, TE 6ms, 1024 vector size, 2000khz bandwidth, 25x30x30mm FOV, and 384 number of averages for GABA measurement. Each RDA data set, generated by Siemens Syngo and its corresponding Siemens MRS raw data were acquired. A constrained nonlinear multivariable numerical optimizer was used for evaluation of optimal weights. SNR was calculated by definition of noise and signal spectra (from 7~8 ppm, NAA, respectively). The metric for optimal weights was defined as maximization of SNR of combined signal. In addition, phase alignment of unsuppressed and suppressed spectra on number of average axis was independently performed using registration with minimum difference metric. All processing was performed by in-house software. There is a significant GABA SNR increase of the optimized weight averaging and phase aligned mean. From simple mean of nSNR weight, average GABA SNR of phase aligned method, optimized weight method, and the combined method increased by 28.3%, 19.8%, and 62.8% respectively.
It was found that the optimal weights combination and phase alignment of individual spectra enhance the quality of MR spectroscopy. Suppression of noise for more correct alignment and optimization would be a further study.
Kim, N,
Sung, Y,
Prescot, A,
Lyoo, I,
Renshaw, P,
SNR-enhancement Using Numerical Optimizer and Phase Alignment for MEGAPRESS. Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL.
http://archive.rsna.org/2010/9012862.html