Spectral Resolution

Spectral resolution essentially refers to the ability to distinguish adjacent peaks in a spectrum of individual peaks. The resonance frequency <w of a specific metabolite is defined by the static magnetic field B0 and a shielding factor o, defined by the geometric structure of the molecule shielding the nuclei from the external magnetic field and referred to as chemical shift. The chemical shift is field-independent, for every metabolite characteristic unit measured as parts per million (ppm). The resonance frequency co(o, B0) of a specific metabolite is defined by:

where y is the gyromagnetic ratio of the nucleus studied. The spectral distance A<w between two peaks with chemical shifts o1 and o2 can be expressed using the above equation as:

This means that the spectral distance for two identical metabolites will be twice as much at 3.0 T compared to 1.5 T (Fig. 6.5a).

This gain in spectral resolution with higher field strengths will allow for improved differentiation between peaks that might overlap at lower field strengths. It is necessary to adjust some sequence parameters to make full use of this gain in spectral resolution. The achieved resolution of the raw data acquired is defined by the sampling scheme used. A given number of points are sampled with a specific sampling rate to get a good digital representation of the MR signal. Typically at 1.5 T, a spectroscopy signal is sampled with 2,048 points and a sampling rate of 2,500 Hz, which is equivalent to a sampling interval of 0.4 ms between each point, yielding a total sampling time of 819.2 ms in which the acquisition window is opened. This kind of resolution is recommended at

Frequency (HlJ

Frequency (HlJ

Fig. 6.5. The frequency resolution between individual peaks increases as a proportion of field strength. This means that the spectral distance Aa> for two identical metabolites will be twice as much at 3.0 T compared to 1.5 T (a) and therefore in abnormal tissue it is possible to resolve different close resonance signals such as glycine at 3.56 ppm from myo-inositol at 3.55 ppm and to achieve a better detection of scyllo-inositol (3.35 ppm) (b)

1.5 T to cover most of the spectral information, and sample long enough to see the signal decay down to the noise level. To cover exactly the same spectral information at 3.0 T, it is necessary to double the sampling rate to 5,000 Hz, which for these settings is equivalent to a sampling interval of 0.2 ms between points. At a constant number of 2,084 points sampled the acquisition window is open for an overall 409.6 ms of sampling time. This sampling window can be so short, especially for phantom experiments, that a significant amount of signal is left at the end of the sampling interval. This can cause artefacts, known as ringing artefacts, in the reconstructed spectrum.

Even though the above considerations are usually of minor importance to the clinical user, they show that a direct one-to-one comparison of MR spectroscopy experiments at different field strengths can be misleading. If exactly the same parameters for sampling rate and number of points are used, the quality of the spectrum acquired at higher field strength is always degraded. If the parameters are adjusted the comparison is no longer one-to-one anymore.

Spectral resolution also depends on the attainable linewidths, which are a function of field-dependent T2 relaxation times and field homogeneity. However, higher-field MR scanners can improve the resolution between peaks as shown above, allowing a more accurate identification and quantification of each metabo

a a.o 7,5 7,0 6.5

lite [16-19, 21]. Despite shorter T2 relaxation times and increased field inhomogeneity, the chemical shift doubling at 3 T yields better spectral resolution. At 1.5 T, the quantification of NAA, Cho, Cr, Lac, and other metabolites such as ml has been feasible [1-4], while Glu and Gln are closely coupled and present extensive spectral overlap at this field intensity, making peak assignments and quantitative measurements difficult and subject to considerable uncertainty. At 3.0 T the increase in chemical shift is reflected, for instance, in improved baseline separation of Cho and Cr, which are only 0.2 ppm apart, and in slightly better resolution of Glu/Gln region, between 2.05 and 2.5 ppm. Furthermore the presence of abnormal metabolites such as Phe at 7.36 ppm (Fig. 6.6) or the differentiation of glycine at 3.56 ppm from myo-inositol at 3.55 ppm (Fig. 6.5b) can be confirmed with more confidence.

Significantly improved spectral resolution was also demonstrated in the quantification of J-coupled metabolites, such as glutamate, glutamine and GABA and the detection of glucose at 3.48 ppm and 5.23 ppm, without using glucose infusion [16,17, 21, 25].

The previous paragraphs have shown that spectral quality is always a compromise between SNR and the spatial, spectral and temporal resolutions. The relations cannot be expressed in a simple formula, and even though systems with higher field strength provide the required basis for a trend to higher resolutions or

-i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|— b 0,0 7.5 7.0 6.5

Fnequency (ppm)

Fig. 6.6. Detection of phenylalanine (Phe) at 7.36 ppm in a patient affected by phenylketonuria (PKU). With respect to the MRS study conducted with the 1.5 T system (a) to detectthe Phe signal it is mandatoryto useaVOI of32 cc andto comparetwo signals in a follow-up therapeutic study. At 3 T, the Phe signal is detectable and can be analysed due to the increased SNR and the better spatial resolution. In fact, use of a smaller VOI (8 cc) reduced the line-broadening effects, allowing Phe to be better resolved with respect to histidineandhomocarnosineat 7.05,7.8 and 8.02 ppm, to the amide proton of NAA at 7.9 ppm and due to the reduction in the macromolecular contributions (MM) at 7.3 ppm

SNR, experimental settings have to be carefully weighted against each other to provide the optimum raw data required for further analysis.

Unraveling Alzheimers Disease

Unraveling Alzheimers Disease

I leave absolutely nothing out! Everything that I learned about Alzheimer’s I share with you. This is the most comprehensive report on Alzheimer’s you will ever read. No stone is left unturned in this comprehensive report.

Get My Free Ebook


Post a comment