Design of broadband RF pulses with polynomial-phase response

The achievable bandwidth of common linear-phase RF pulses is limited by the maximum feasible B1 amplitude of the MR system. It has been shown previously, that this limitation can be circumvented by overlaying a quadratic phase in the frequency domain, which spreads the power across the pulse duration. Quadratic-phase RF pulses are near optimal in terms of achieving minimal B1max. In this work, it is demonstrated that further B1max reduction can be achieved by combining quadratic with higher-order polynomial-phase functions. RF pulses with a phase response up to tenth order were designed using the Shinnar-Le Roux transformation, yielding considerable increases in bandwidth and selectivity as compared to pure quadratic-phase pulses. These benefits are studied for a range of pulse specifications and demonstrated experimentally. For B1max = 20 microT and a pulse duration of 2.1 ms, it was possible to increase the bandwidth from 3.1 kHz for linear and 3.8 kHz for a quadratic to 9.9 kHz for a polynomial-phase pulse.     

The rα-structures of partially oriented 1,2,4,5-and 1,2,3,4-tetrachlorobenzene determined from their proton magnetic resonance spectra with natural abundance 13C-satellites

By recording two spectra of the same molecule in different nematic liquid crystals, and analysing the data simultaneously, information has been obtained on the molecular structures of 1,2,4,5- (1) and 1,2,3,4-tetrachlorobenzene (2). Carbon–hydrogen and carbon–carbon internuclear distance ratios, as well as bond lengths and bond angles, were computed and corrected for harmonic vibrations. Only part of the molecular structure of 2 could be derived with reasonable precision; the reasons for this limitation are discussed. The spectra of the two isomers were also analysed in isotropic solvents in order to determine the indirect coupling constants. One- and two-bond ¹³C isotope effects on the chemical shifts of the protons were observed. The solvent effects of the different nematic liquid crystals on the structure were found to be small compared with the experimental error.     

Linear Response Equilibrium versus echo-planar encoding for fast high-spatial resolution 3D chemical shift imaging

In this work Linear Response Equilibrium (LRE) and Echo-planar spectroscopic imaging (EPSI) are compared in terms of sensitivity per unit time and power deposition. In addition an extended dual repetition time scheme to generate broad stopbands for improved inherent water suppression in LRE is presented. The feasibility of LRE and EPSI for assessing cholesterol esters in human carotid plaques with high spatial resolution of 1.95×1.15×1.15 mm(3) on a clinical 3T MR system is demonstrated. In simulations and phantom experiments it is shown that LRE has comparable but lower sensitivity per unit time relative to EPSI despite stronger signal generated. This relates to the lower sampling efficiency in LRE relative to EPSI as a result of limited gradient performance on clinical MR systems. At the same time, power deposition of LRE is significantly reduced compared to EPSI making it an interesting niche application for in vivo high field spectroscopic imaging of metabolites within a limited bandwidth.     

Tissue characterization of brain tumors during and after pion radiation therapy

Negative Pi-mesons (pions) are applied at the Paul Scherrer Institute in the radiotherapy of highly malignant gliomas using a dose escalation program. The therapy effects of 7 randomly selected patients were followed up by 62 MRI examinations. The quantification of the effects is based on the relaxation times T1 and T2, which are acquired by a new designed multi-echo multiple saturation recovery imaging technique. As a summary of the results, roughly two reaction types are observed. For both types the relaxation times increase up to two to three months after the radiation therapy. Then in one type (two patients) the T1 and T2 values of the tumors, and of the edemas surrounding the tumors, further increase, indicating an unfavorable prognosis. In the other type (five patients) the relaxation times drop down towards, or even below, their initial values, reflecting the onset of the reparation processes in the tissue. This later behaviour reflects an at least temporary control of the disease; that is, the short term prognosis for these patients is more favorable. It further can be concluded, with respect to our MR parameters, that the radiotolerance of healthy brain tissue is much higher than that of malignant glioma tissue, despite the fact that these tumors are very seldom definitively radiosensible.     

FID-acquired-echos (FAcE): a short echo time imaging method for flow artefact suppression.

The FID-Acquired-Echo sequence (FAcE) is a magnetic resonance imaging technique using fractional-echo acquisitions, with sequential separate sampling of the right and left k-space half planes. It reduces the minimal echo times by about a factor of two, compared to conventional full-(gradient)-echo sampling schemes. With this sequence, implemented on a commercial 1.5 Tesla whole body system, high resolution images are acquired with typical echo times between 3 and 4.5 msec. Using short echo times the signal dephasing caused by velocity and higher order spin motion is reduced. Further, due to the modified sampling scheme, the sequence exhibits, for triggered studies, partially a compensation of motion-induced phase shifts in the frequency-encoding direction. Thus, the sequence offers an alternative means for the reduction of motion-induced image artefacts to the use of flow compensating gradients, which usually makes a sequence more sensitive to higher order motion and introduces further eddy currents. Besides potential application for imaging of nuclei and tissues with short T2 relaxation times, and non-ECG-triggered in-flow angiography, the main application seems to be triggered-phase contrast imaging with focus on quantitation of blood flow. Its usefulness is largest in cases with irregular flow patterns, where considerable in-plane flow occurs.     

Quantitative abdominal aortic flow measurements at controlled levels of ergometer exercise

Measuring the exercise-induced flow changes in the arteries of the body is a major challenge. The use of quantitative MR flow measurements for this purpose is hampered by movement artifacts and ECG triggering problems. To quantify exercise-induced flow changes in the abdominal aorta, we applied a fast hybrid phase contrast sequence with K-space segmentation and echo planar imaging readouts during a 12 heart beat, single breathhold post exercise scanning window after ergometer exercise in nine volunteers. Central k-space was acquired first. The changes in heart rate throughout the scanning window were quantified. The mean decrease in heart rate after six heart beats post exercise was less than 4% and less than 14% after 11 heart beats indicating that the exercise state was very well represented during the acquisition of central k-space. Abdominal aortic flow increased from 1.4+/-0.3 l/min at rest to 7.9+/-1.1 l/min at 131 watt. Retrograde flow reached a maximum value of 1.2 l/min at rest, and lasted 140 ms on average. Only for one out of the nine volunteers was there any retrograde flow present during exercise (at 33 watt and 65 watt exercise). It was concluded that retrograde flow patterns in the abdominal aorta associated with oscillating wall shear stresses and development of atherosclerosis disappeared with increasing levels of exercise. The feasibility of using fast quantitative phase contrast measurements during a post exercise scanning window to represent controlled exercise levels was demonstrated.     

A multi-sample 94 GHz dissolution dynamic-nuclear-polarization system

We describe the design and initial performance results of a multi-sample dissolution dynamic-nuclear-polarization (DNP) polarizer based on a Helium-temperature NMR cryostat for use in a wide-bore NMR magnet with a room-temperature bore. The system is designed to accommodate up to six samples in a revolver-style sample changer that allows changing samples at liquid-Helium temperature and at pressures ranging from ambient pressure down to 1 mbar. The multi-sample setup is motivated by the desire to do repetitive in vivo measurements and to characterize the DNP process by investigating samples of different chemical composition. The system can be loaded with up to six samples simultaneously to reduce sample loading and unloading. Therefore, series of experiments can be carried out faster and more reliably. The DNP probe contains an oversized microwave cavity and includes EPR and NMR capabilities for monitoring the DNP process. In the solid state, DNP enhancements corresponding to ～45% polarization for [1-(13)C]pyruvic acid with a trityl radical have been measured. In the initial liquid-state acquisition experiments described here, the polarization was found to be ～13%, corresponding to an enhancement factor exceeding 16,000 relative to thermal polarization at 9.4 T and ambient temperature.     

Effects of vigabatrin intake on brain GABA activity as monitored by spectrally edited magnetic resonance spectroscopy and positron emission tomography

A deficit in gamma-aminobutyric acid (GABA) levels in the brain or the cerebrospinal fluid (CSF) is found in many epilepsy patients. Frequency and severity of seizures may be reduced by treatment with GABA increasing medicaments as e.g. vigabatrin, an irreversible inhibitor of GABA-transaminase. For a better understanding of the associated effects, healthy volunteers were examined with magnetic resonance spectroscopy (MRS) and positron emission tomography (PET) before and after intake of different doses of vigabatrin. For the MRS examinations, a dedicated localized spectral editing method was developed to determine GABA levels. The 11C-flumazenil (FMZ)-PET protocol allowed determination of GABA-A receptor binding. The results show a clear and dose-dependent increase in the brain GABA levels after the medication period as compared to the baseline values. The GABA-A receptor binding, on the other hand, did not change significantly.     

In vivo 1H NMR spectroscopy of individual human brain metabolites at moderate field strengths

This article reviews spectral editing techniques for in vivo ¹H NMR spectroscopy of human brain tissue at moderate field strengths of 1.5–3 Tesla. Various aspects of ¹H NMR spectroscopy are discussed with regard to in vivo applications. The parameter set [δ, J, n] (δ being the relative chemical shift, J the scalar coupling constant and n the number of coupled spins) is used to characterize the spin systems under investigation and to classify the editing techniques that are used in in vivo ¹H NMR spectroscopy.     

Equi-ripple design of quadratic-phase RF pulses

An improved strategy for the design of quadratic-phase RF pulses with high selectivity and broad bandwidths using the Shinnar-Le Roux (SLR) transformation is proposed. Unlike previous implementations, the required quadratic-phase finite impulse response (FIR) filters are generated using the complex Remez exchange algorithm, which ensures an equi-ripple deviation from the ideal response function. It is argued analytically that quadratic-phase pulses are near-optimal in terms of minimising the B1-amplitude for a given bandwidth and flip angle. Furthermore, several parameter relations are derived, providing practical design guidelines. The effectiveness of the proposed design method is demonstrated by examples of excitation and saturation pulses applied in vitro and in vivo.