No abstract

No abstract

A new two-scan method for localized 1 H in vivo NMR spectroscopy (MRS) without water suppression (WS) is described. In one of the scans, two chemical shift selective 180°pulses are applied prior to a standard localization sequence to invert all metabolite signals upfield and downfield from water, which remains unaffected. The difference spectrum records the metabolites whereas water and accompanying gradient induced artifacts are widely suppressed. The method was implemented on a 4.7-T system using point resolved spectroscopy with a short echo time of 18 ms. Phantom measurements proved the feasibility of absolute quantification using water as an internal reference. Measurements on healthy rat brain yielded comparable spectrum quality as measurements with water presaturation. The method does not require additional adjustments or sophisticated data postprocessing and scales favorably with increasing B 0 field. Therefore, the method should be useful for 1 H MRS without WS. Although the two-step method doubles the minimum total measurement time, it may also be of interest for spectroscopic imaging ( For more than one decade an optimized suppression of the predominant water signal had been considered a prerequisite for high-quality localized in vivo 1 H NMR spectroscopy (MRS) or spectroscopic imaging (SI) (1-4). However, there are distinct reasons to perform MRS measurements without water suppression (WS). First, the water signal can be used as a reference signal for absolute quantification (5) and/or for correction procedures such as phase or eddy current corrections (6). Second, WS techniques may cause magnetization transfer effects to metabolite signals (7,8), leading to systematic quantification errors. Third, adjustments associated with WS techniques can be avoided and the RF power deposition can be reduced. Fourth, it may be possible to detect signals with rather small chemical shift differences to water.Following the early publications of Hurd et al. (9) and Kreis and Boesch (10) in 1998, various methods have been proposed to realize in vivo 1 H MRS or SI without WS (11-16). The dynamic range of the digitized signal is large enough to avoid additional digital noise if an analog-todigital converter (ADC) with high digital resolution is used and digital filtering using oversampling is applied. A 16-bit ADC and digital filtering, which yields additional 2-to 4-bit resolution, proved to be adequate for typical singlevoxel MRS measurements because the analog noise, produced by the RF coil and/or the sample, covers at least 3-4 bits after digitization. Only if spectroscopic imaging of a large 2D or 3D volume is performed, may a larger dynamic range, i.e., an ADC with higher digital resolution, be necessary.The second main technical problem of MRS without WS, the appearance of gradient-induced sidebands of the water signal resulting from a frequency modulation of the water signal, is more difficult to overcome. Different solutions have been proposed considering the origin and the properties of the gradient-induced ar...

This article examines the way in which microscopic tissue parameters affect the signal attenuation of diffusion-weighted MR experiments. The influence of transmembrane water flux on the signal decay is emphasized using the Kä rger equations, which are modified with respect to the cellular boundary restrictions for intra-and extracellular diffusion. This analytical approach is extensively compared to Monte-Carlo simulations for a tissue model consisting of two compartments. It is shown that diffusion-weighted MR methods provide a unique tool for estimation of the intracellular exchange time. Restrictions of applicability to in vivo data are examined. It is shown that the intracellular exchange time strongly depends on the size of a cell, leading to an apparent diffusion time dependence for in vivo data. Hence, an analytical model of a two-compartment system with an averaged exchange time is inadequate for the interpretation of signal curves measured in vivo over large ranges of b-values. During recent years, diffusion-weighted nuclear magnetic resonance methods (DW-NMR) have attracted attention in clinical diagnosis for the detection of brain lesions, especially of stroke in its early stage. The image contrast of this method is based on the differences in the apparent diffusion coefficients (ADC) of water in healthy and infarcted brain regions. The ADC reflects diffusion properties within tissue and depends on many physiological parameters such as volume fractions, extracellular tortuosity, intracellular restrictions, membrane permeability, and active processes across membranes, relaxation rates, or anisotropic morphology etc. Hence, the interpretation of diffusionweighted data is a complex multiparameter problem and it is crucial to extract the parameters specified above from the ADC in order to gain a better characterization of the physiological state of the tissue. Thus, much effort has been invested in the analytical description of diffusion effects on the NMR signal (see, for example, Refs. 1-4). However, any proposed mathematical model only provides an approximation with vague limits of applicability, in particular because many model input parameters, such as the intracellular diffusion coefficient, the tortuosity factor, or the exchange rates of water across cellular membranes are still not clearly known. Furthermore, their effect on the signal attenuation depends on the experimental parameters, such as diffusion time or gradient strength. Therefore, analytical models have often been compared to numerical simulations (2,3,5). Many experiments performed in vitro and in vivo showed that the diffusion attenuation is not monoexponential when measured over a wide range of b-values (3,6 -14). This behavior presumably reflects tissue compartmentalization and may provide structural information. Hence, deviations from monoexponentiality can be used to separate the influence of tissue parameters on the attenuation and may therefore be exploited when testing analytical models. This is the approach set forth in this a...

The use of water suppression for in vivo proton MR spectroscopy diminishes the signal intensities from resonances that undergo magnetization exchange with water, particularly those downfield of water. To investigate these exchangeable resonances, an inversion transfer experiment was performed using the metabolite cycling technique for non-watersuppressed MR spectroscopy from a large brain voxel in 11 healthy volunteers at 3.0 T. The exchange rates of the most prominent peaks downfield of water were found to range from 0.5 to 8.9 s 21, while the T 1 relaxation times in absence of exchange were found to range from 175 to 525 ms. The use of water suppression for in vivo proton MR spectroscopy (MRS) arose as a solution for difficulties with the limited dynamic range of analog-to-digital converters and baseline distortions caused by the several orders of magnitude difference in concentration between water and metabolite protons. However, the use of water suppression can also diminish signal from metabolites of interest including metabolites that exchange magnetization with water, such as those containing amide moieties, resulting in spectra with decreased information content. Acquiring spectra without water suppression enables automatic phase, frequency and eddy current correction of individual traces, as well as offers an internal reference signal for absolute quantification, without the requirement of additional measurements. Unfortunately, these advantages are hindered by the presence of gradient coil vibration artifacts, also known as ''sidebands,'' caused by temporal fluctuations in the local amplitude of static (polarizing) field (B 0 ) following gradient switching (1).Several experimental and post-acquisition processing methods have been proposed to eliminate sidebands from MR spectra acquired with clinical systems. Experimental methods proposed to date can be roughly categorized as (a) selective dephasing of water, (b) gradient cycling, (c) metabolite cycling, (d) B 0 compensation, or (e) sideband subtraction. Selective dephasing gradients (such as those found in the WATERGATE (WATER suppression by GRAdient Tailored Excitation) sequence) can be incorporated into the evolution period of a localization sequence to selectively dephase water before detection, thus minimizing time for magnetization exchange; however, selective pulses with sharp transition bands are relatively long and, hence, do not allow for short echo times (TEs) at clinical field strengths (2,3). (b) Gradient cycling involves alternately inverting the direction of the gradient pulses used in localization (4-7), since the phase of the sidebands depends on the sign of the gradients. Though the sideband artifacts are significantly reduced by this method, the extent of the reduction depends on the hardware details of the scanner and the sequence, and the resulting spectra contain a large water peak whose tail overlaps the metabolite regions necessitating further processing. (c) The metabolite cycling technique uses a frequency selective inversion p...