Phase-encoding strategies for optimal spatial resolution and T1 accuracy in 3D Look-Locker imaging

The Look-Locker (LL) imaging method provides an accurate and efficient approach for mapping the spin-lattice relaxation time, T(1). However, the same recovery of signal during LL image acquisition required to estimate T(1) also results in unwanted modulation of k-space. This is particularly problematic with 3D LL imaging as the number of phase-encoding steps during the recovery interval (e.g., 16) increases in an effort to reduce imaging times. This modulation of k-space has the effect of introducing a point spread function (PSF), which can lead to either image blurring (if the earlier tip angles are assigned to the centre of k-space) or edge enhancement (if the earlier tip angles are assigned to the edges of k-space), thus corrupting T(1) estimation, particularly for small objects. In this study, the PSF and its effect on the acquired images for four different interleaved phase-encode schemes (centric-in, centric-out, sequential and hybrid-sequential) are simulated for a range of T(1), tip angle and 3D LL acquisition parameters expected in practice. It is shown by simulation and confirmed experimentally in phantoms that a hybrid sequential phase-encoding scheme reduces image blurring while maintaining T(1) accuracy ( approximately 2%) and precision (2%) over a range of object sizes down to 2 pixels (2 mm).     

SAR reduced pulse sequences

Three techniques were considered for reducing the RF (radiofrequency) power deposition in the body while maintaining scan time efficiency: reducing the RF peak amplitude while increasing the pulse width, substituting gradient echoes for spin echoes, and reducing the flip angle of the phase reversal pulse. The use of gradient echoes was found to be the most efficient means to reduce the power delivered to the patient and to obtain rapid data acquisition. The effect upon SAR (specific absorption rate) and SNR (signal-to-noise ratio) was demonstrated on a phantom when the phase reversal pulse was reduced from the standard 180 degrees to 90 degrees. Data in the body indicated a fairly constant SNR down to a refocusing flip angle between 110 degrees and 135 degrees. An initial clinical evaluation was performed at three institutions using the method of reducing the flip angle of the phase reversal pulse. The scan with theta = 120 degrees was rated by readers in a blinded study as having acceptable diagnostic image quality while the 135 degrees scan had comparable image quality to a conventional 90 degrees - 180 degrees pulse sequence. The use of reduced phase reversal pulses was seen as an efficient protocol to obtain T1-weighted images at rapid data rates while reducing the power delivered to the body by about 40%.     

Maximizing signal-to-noise and contrast-to-noise ratios in FLASH imaging

This paper presents an analysis of signal-to-noise and contrast-to-noise ratios from small tip angle, gradient reversal (FLASH) imaging. Analytic and numerical techniques are used to determine the delay times and tip angles that maximize signal-to-noise per unit time from a single tissue. Similar procedures are used to determine the delay times and tip angles that maximize both T1-induced and T-2*-induced contrast-to-noise per unit time for a pair of tissues as a function of tissue characteristics and pulse sequence sampling times. The advantage of optimized FLASH imaging over optimized spin-echo imaging is quantitated by comparing signal-to-noise and contrast-to-noise ratios per unit time from the two sequences. Images are used to confirm these numerical results, to compare noise levels resulting from gradient reversals versus 180 degrees rephasing pulses and to assess the possible adverse effects of static magnetic field inhomogeneities on FLASH imaging.     

Sodium 3-D MRI of the human torso using a volume coil

Sodium MR imaging is considered to provide clinically important information about the human body that is not achievable by hydrogen-based approaches. However, due to the low natural abundance in biological tissues, sodium signals usually lead to low spatial resolution, low SNR, and long acquisition times compared to conventional 1H imaging, even using well-adapted surface coils. For our study, a volume coil was designed with nearly homogeneous excitation/receive characteristics and a suitable geometry fitting the human torso. A sufficient penetration throughout the entire thorax, abdomen, or pelvis is provided allowing for sodium imaging of the kidneys, the liver with gall bladder, or the myocardium. All measurements were performed on a 1.5 T whole body scanner using a spoiled 3-D gradient echo sequence. Imaging parameters TE, TR, and readout bandwidth were optimized for sensitive recording of the sodium component with slow transverse relaxation. Nonselective RF excitation pulses with a duration of 2.5 ms and rectangular shape were applied to avoid SAR problems. Narrow receiver bandwidth and excitation near the Ernst angle provided clinically practicable examinations with measuring times of less than 15 min at a spatial resolution of 8 x 8 x 8 mm3. Under these conditions, SNR of 11 for the kidneys and vertebral disks, 9 for the spinal canal, and 6 for the liver was achieved. A special 3-D spin echo sequence was used to determine T2, times which resulted to 15.3 +/- 1.1 ms for liver, 27.7 +/- 7.2 ms for kidneys, and 24.0 +/- 4.7 ms for the content of the spinal canal.     

T1 measurements incorporating flip angle calibration and correction in vivo

In this work, we propose a variable FA method that combines in vivo flip angle (FA) calibration and correction with a short TR variable FA approach for a fast and accurate T(1) mapping. The precision T(1)s measured across a uniform milk phantom is estimated to be 2.65% using the conventional (slow) inversion recovery (IR) method and 28.5% for the variable FA method without FA correction, and 2.2% when FA correction is included. These results demonstrate that the sensitivity of the variable FA method to RF nonuniformities can be dramatically reduced when these nonuniformities are directly measured and corrected. The acquisition time for this approach decreases to 10 min from 85 min for the conventional IR method. In addition, we report that the averaged T(1)s measured from five normal subjects are 900 +/- 3 ms, 1337 +/- 8 ms and 2180 +/- 25 ms in white matter (WM), gray matter (GM) and cerebral spinal fluid (CSF) using the variable flip angle method with FA correction at 3 T, respectively. These results are consistent with previously reported values obtained with much longer acquisition times. The method reduces the total scan time for whole brain T(1) mapping, including FA measurement and calibration, to approximately 6 min. The novelty of this method lies in the in vivo calibration and the correction of the FAs, thereby allowing a rapid and accurate T(1) mapping at high field for many applications.     

Comparative study of fast MR imaging: quantitative analysis on image quality and efficiency among various time frames and contrast behaviors

The purpose of this study is to quantitatively compare the image quality and efficiency provided by widely available fast MR imaging pulse sequences. A composite phantom with various T1 and T2 values and subjected to periodic motion was imaged at 1.5 T. The fast MRI sequences evaluated included fast spin-echo (FSE), single shot fast spin-echo (SSFSE), echo-planar imaging (EPI), multi-slice gradient recalled (MPGR), fast MPGR (FMPGR), and fast multi-slice spoiled gradient echo (FMPSPGR). T1-weighted (T1WI), T2-weighted (T2WI), proton-density-weighted (PDWI), and T2*-weighted (T2*WI) images were evaluated in breath-hold and non-breath-hold time frames. Analysis included measurement of image signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), nonuniformity, ghosting ratio, SNR per unit time and CNR per unit time. Among fast T2WI sequences, FSE with breath-hold time frame resulted in the highest image quality and in superior SNR and CNR efficiency by a factor of 5 or 6 as compared with conventional spin echo sequence. Among fast T1WI sequences, FMPGR and FMPSPGR both with non-breath-hold time frame produced the highest image quality and SNR and CNR efficiency by a factor of greater than 5 as compared with conventional spin echo. Among fast PDWI and T2*WI sequences, FSE produced the highest SNR and CNR, and was maximally efficient with a factors of greater than 6 as compared with conventional spin echo.     

A comparison and evaluation of reduced-FOV methods for multi-slice 7 T human imaging

Eight different reduced field-of-view (FOV) MRI techniques suitable for high field human imaging were implemented, optimized, and evaluated at 7 T. These included selective Inner-Volume Imaging (IVI) based methods, and Outer-Volume Suppression (OVS) techniques, some of which were previously unexplored at ultra-high fields. Design considerations included use of selective composite excitation and adiabatic refocusing radio-frequency (RF) pulses to address B1 inhomogeneities, twice-refocused spin echo techniques, frequency-modulated pulses to sharply define suppressed regions, and pulse sequence designs to improve SNR in multi-slice scans. The different methods were quantitatively compared in phantoms and in vivo human brain images to provide measurements of relative signal to noise ratio (SNR), power deposition (specific absorption rate, SAR), suppression of signal, artifact strength and prevalence, and general image quality. Multi-slice signal losses in out-of-slice locations were simulated for IVI methods, and then measured experimentally across a range of slice numbers. Corrections for B1 nonuniformities demonstrated an improved SNR and a reduction in artifact power in the reduced-FOV, but produced an elevated SAR. Multi-slice sequences with reordering of pulses in traditional and twice-refocused IVI techniques demonstrated an improved SNR compared to conventional methods. The combined results provide a basis for use of reduced-FOV techniques for human imaging localized to a small FOV at 7 T.     

Steady-state free precession with hyperpolarized 3He: experiments and theory

The magnetization response of hyperpolarized 3He gas to a steady-state free precession (SSFP) sequence was simulated using matrix product operators. The simulations included the effects of flip angle (alpha), sequence timings, resonant frequency, gas diffusion coefficient, imaging gradients, T1 and T2. Experiments performed at 1.5 T, on gas phantoms and with healthy human subjects, confirm the predicted theory, and indicate increased SNR with SSFP through use of higher flip angles when compared to optimized spoiled gradient echo (SPGR). Simulations and experiments show some compromise to the SNR and some point spread function broadening at high alpha due to the incomplete refocusing of transverse magnetization, caused by diffusion dephasing from the readout gradient. Mixing of gas polarization levels by diffusion between slices is also identified as a source of signal loss in SSFP at higher alpha through incomplete refocusing. Nevertheless, in the sample experiments, a SSFP sequence with an optimized flip angle of alpha=20 degrees, and 128 sequential phase encoding views, showed a higher SNR when compared to SPGR (alpha=7.2 degrees) with the same bandwidth. Some of the gas sample experiments demonstrated a transient signal response that deviates from theory in the initial phase. This was identified as being caused by radiation damping interactions between the large initial transverse magnetization and the high quality factor (Q=250) birdcage resonator. In 3He NMR experiments, performed without imaging gradients, diffusion dephasing can be mitigated, and the effective T2 is relatively long (1 s). Under these circumstances the SSFP sequence behaves like a CPMG sequence with sinalpha/2 weighting of SNR. Experiments and simulations were also performed to characterize the off-resonance behaviour of the SSFP HP 3He signal. Characteristic banding artifacts due to off-resonance harmonic beating were observed in some of the in vivo SSFP images, for instance in axial slices close to the diaphragm where B0 inhomogeneity is highest. Despite these artifacts, a higher SNR was observed with SSFP in vivo when compared to the SPGR sequence. The trends predicted by theory of increasing SSFP SNR with increasing flip angle were observed in the range alpha=10-20 degrees without compromise to image quality through blurring caused by excessive k-space filtering.     

Evaluation of Enhancement Effects as a Function of the Molarity of Gd-Based Contrast Media at 3.0 and 1.5 T: Based on the T1 Effective Pulse Sequence Parameter

The purpose of this study was to evaluate the alterations of diluted molarity of contrast media to emit the maximum signal intensity by changing the parameters of pulse sequences. The phantom was developed by diluting the magnetic resonance imaging (MRI) T1 contrast medium. The phantom images were obtained by 1.5 and 3.0 T MRI systems. We conducted Pearson’s analysis to reveal the correlation of the signal-to-noise ratio (SNR)_90%, the change of the concentration range of the contrast media which shows over 90% SNR, with changing the parameters of T1 effect pulse sequences in both 1.5 and 3.0 T imaging. As the flip angle increased, the SNR increased for all contrast media in magnetization-prepared rapid gradient echo and two-dimensional fast low angle shot pulse sequences at 1.5 and 3.0 T. Although the SNR increased until 30°, the SNR was almost the same over 30° in volumetric interpolated breath-hold examination at 1.5 and 3.0 T. The minimum contrast molarity of the representing SNR_90% was decreased according to the increasing time to repeat in spin echo. The present study revealed that the high concentration technique of contrast media on three pulse sequences (VIBE, MPRAGE, and 2D FLASH) could be useful to obtain images with better SNR.     

Spin-lattice relaxation and a fast T1-map acquisition method in MRI with transient-state magnetization

The magnetization under the spin-lattice relaxation and the nuclear magnetic resonance radiofrequency (RF) pulses is calculated for a signal RF pulse train and for a sequence of multiple RF pulse-trains. It is assumed that the transverse magnetization is zero when each RF pulse is applied. The result expressions can be grouped into two terms: a decay term, which is proportional to the initial magnetization M0, and a recovery term, which has no M0 dependence but strongly depends on the spin-lattice relaxation and the equilibrium magnetization Meq. In magnetic resonance pulse sequences using magnetization in transient state, the recovery term produces artifacts and can seriously degrade the function of the preparation sequence for slice selection, contrast weighting, phase encoding, etc. This work shows that the detrimental effect can be removed by signal averaging in an eliminative fashion. A novel fast data acquisition method for constructing the spin-lattice relaxation (T1) map is introduced. The method has two features: (i) By using eliminative averaging, the curve to fit the T1 value is a decay exponential function rather than a recovery one as in conventional techniques; therefore, the measurement of Meq is not required and the result is less susceptible to the accuracy of the inversion RF pulse. (ii) The decay exponential curve is sampled by using a sequence of multiple pulse-trains. An image is reconstructed from each train and represents a sample point of the curve. Hence a single imaging sequence can yield multiple sample points needed for fitting the T1 value in contrast to conventional techniques that require repeating the imaging sequence for various delay values but obtain only one sample point from each repetition.     

Rapid MRI method for mapping the longitudinal relaxation time

A novel method for mapping the longitudinal relaxation time in a clinically acceptable time is developed based on a recent proposal [J.-J. Hsu, I.J. Lowe, Spin-lattice relaxation and a fast T1-map acquisition method in MRI with transient-state magnetization, J. Magn. Reson. 169 (2004) 270-278] and the speed of the spiral pulse sequence. The method acquires multiple curve-fitting samples with one RF pulse train. It does not require RF pulses of specific flip angles (e.g., 90 degrees or 180 degrees ), nor are the long recovery waiting time and the measurement of the magnetization at thermal equilibrium needed. Given the value of the flip angle, the curve fitting is semi-logarithmic and not computationally intensive. On a heterogeneous phantom, the average percentage difference between measurements of the present method and those of an inversion-recovery method is below 2.7%. In mapping the human brain, the present method, for example, can obtain four curve-fitting samples for five 128 x 128 slices in less than 3.2s and the results are in agreement with other studies in the literature.     

Simultaneous measurement of temperature and velocity maps by inversion recovery tagging method

A new method, called the inversion recovery (IR) tagging method, for simultaneous measurement of temperature and velocity maps of flowing fluid has been developed. The present method employs a set of tagging pulses which acts as an inversion pulse of the conventional IR method, based on the temperature dependence of the spin-lattice relaxation of water proton in a fluid, and has the advantage of being able to compensate the reduction of the NMR signal intensity due to flow motion and to reduce the total time to measure these maps. First, the accuracy of the temperature measurement of stagnant doped water in a differentially heated cell using the conventional IR method, as the basic sequence of the IR tagging method, has been evaluated. The accuracy was within 10% of the temperature difference DeltaT = 17.2 degrees C and the measurable temperature resolution was within +/-0.5 degrees C. Then temperature and velocity maps of the flowing doped-water through a cooled pipe were measured simultaneously by the IR tagging method, and the accuracy of temperature measurement was evaluated. The accuracy obtained using the present method was within 15% of the temperature difference DeltaT = 15 degrees C.     

A closer look into DESIRE for NMR microscopy

The major challenge of nuclear magnetic resonance (NMR) microscopy at a spatial resolution of a few micrometers is to obtain a sufficiently high signal-to-noise-ratio (SNR) within a reasonable measurement time. As a particular difficulty, molecular self-diffusion poses a serious limitation to true spatial resolution and SNR if conventional Fourier encoding techniques are used. Opposed to that, the alternative DESIRE (Diffusion Enhancement of SIgnal and REsolution) approach to NMR microscopy utilises diffusion to increase the SNR. Being a real-space imaging method, spatial localisation is accomplished by saturation pulses while diffusion continuously replaces the saturated by unsaturated spins. For this technique a signal enhancement of up to three orders of magnitude has been predicted and initial experimental data have provided a proof of principle. In the present work, a detailed investigation of one-dimensional (1D) DESIRE is presented including simulations of a real implementation of the method, a quantitative experimental analysis, and basic 1D imaging. The simulations reveal the importance and provide the means of ensuring the true spatial resolution for this particular way of localisation, enable the selection of useful experimental parameters, and predict the specific image contrast to be expected around barriers restricting diffusion. Experimental data are presented with resolutions down to 3 microm and DESIRE enhancement up to 25 that are in good agreement with the simulation results. In particular, 1D DESIRE imaging in a phantom confirms the expected signal drop close to barriers due to spatially restricted diffusion.     

Development and initial evaluation of 7-T q-ball imaging of the human brain

Diffusion tensor imaging (DTI) noninvasively depicts white matter connectivity in regions where the Gaussian model of diffusion is valid but yields inaccurate results in those where diffusion has a more complex distribution, such as fiber crossings. q-ball imaging (QBI) overcomes this limitation of DTI by more fully characterizing the angular dependence of intravoxel diffusion with larger numbers of diffusion-encoding directional measurements at higher diffusion-weighting factors (b values). However, the former technique results in longer acquisition times and the latter technique results in a lower signal-to-noise ratio (SNR). In this project, we developed specialized 7-T acquisition methods utilizing novel radiofrequency pulses, eight-channel parallel imaging EPI and high-order shimming with a phase-sensitive multichannel B0 field map reconstruction. These methods were applied in initial healthy adult volunteer studies, which demonstrated the feasibility of performing 7-T QBI. Preliminary comparisons of 3 T with 7 T within supratentorial crossing white matter tracts documented a 79.5% SNR increase for b=3000 s/mm2 (P=.0001) and a 38.6% SNR increase for b=6000 s/mm2 (P=.015). With spherical harmonic reconstruction of the q-ball orientation distribution function at b=3000 s/mm2, 7-T QBI allowed for accurate visualization of crossing fiber tracts with fewer diffusion-encoding acquisitions as compared with 3-T QBI. The improvement of 7-T QBI at b factors as high as 6000 s/mm2 resulted in better angular resolution as compared with 3-T QBI for depicting fibers crossing at shallow angles. Although the increased susceptibility effects at 7 T caused problematic distortions near brain-air interfaces at the skull base and posterior fossa, these initial 7-T QBI studies demonstrated excellent quality in much of the supratentorial brain, with significant improvements as compared with 3-T acquisitions in the same individuals.     

Spectroscopic imaging of the brain with phased-array coils at 3.0 T

The goal of this study was to develop and evaluate high-resolution magnetic resonance spectroscopic imaging (MRSI) utilizing the gains in signal-to-noise ratio (SNR) provided by combining higher magnetic field with high-sensitivity phased-array (PA) coils. We investigated the maximum improvement in spatial resolution as small as 0.09 cm(3) for brain MRSI while maintaining adequate SNR and acquisition time. The use of low peak power, dual-band spectral-spatial pulses was also investigated for application to 3 T MRSI of the brain using the body coil for radiofrequency excitation and PA coils for signal reception.     

On neglecting chemical exchange effects when correcting in vivo (31)P MRS data for partial saturation

Signal acquisition in most MRS experiments requires a correction for partial saturation that is commonly based on a single exponential model for T(1) that ignores effects of chemical exchange. We evaluated the errors in (31)P MRS measurements introduced by this approximation in two-, three-, and four-site chemical exchange models under a range of flip-angles and pulse sequence repetition times (T(R)) that provide near-optimum signal-to-noise ratio (SNR). In two-site exchange, such as the creatine-kinase reaction involving phosphocreatine (PCr) and gamma-ATP in human skeletal and cardiac muscle, errors in saturation factors were determined for the progressive saturation method and the dual-angle method of measuring T(1). The analysis shows that these errors are negligible for the progressive saturation method if the observed T(1) is derived from a three-parameter fit of the data. When T(1) is measured with the dual-angle method, errors in saturation factors are less than 5% for all conceivable values of the chemical exchange rate and flip-angles that deliver useful SNR per unit time over the range T(1)/5 < or = T(R) < or = 2T(1). Errors are also less than 5% for three- and four-site exchange when T(R) > or = T(1)(*)/2, the so-called "intrinsic" T(1)'s of the metabolites. The effect of changing metabolite concentrations and chemical exchange rates on observed T(1)'s and saturation corrections was also examined with a three-site chemical exchange model involving ATP, PCr, and inorganic phosphate in skeletal muscle undergoing up to 95% PCr depletion. Although the observed T(1)'s were dependent on metabolite concentrations, errors in saturation corrections for T(R) = 2 s could be kept within 5% for all exchanging metabolites using a simple interpolation of two dual-angle T(1) measurements performed at the start and end of the experiment. Thus, the single-exponential model appears to be reasonably accurate for correcting (31)P MRS data for partial saturation in the presence of chemical exchange. Even in systems where metabolite concentrations change, accurate saturation corrections are possible without much loss in SNR.     

Limited field of view spin echo MR imaging

Several groups have reported using a method of limiting the field of view (FOV) where the slices excited by the 90 and 180 degree pulses are perpendicular. However, only one slice can be excited during each repetition time, so multislice imaging is not possible. We present a modification of this method that allows multislice imaging. The slices excited by the 90 degrees and 180 degrees pulses are at a small angle; the field of view is limited and multislice imaging is possible. The modifications also allow the center of the FOV to be offset to any position. We describe the conditions that yield optimal images for the given FOV, slice thickness, and interslice gap. Representative images demonstrating the features of the technique are presented. The technique can be used to reduce the number of phase-encoding steps resulting in reduced imaging time, or it can be used to increase the spatial resolution without increasing the imaging time.     

Double spin-echo sequence for rapid spectroscopic imaging of hyperpolarized 13C

Dynamic nuclear polarization of metabolically active compounds labeled with (13)C has been introduced as a means for imaging metabolic processes in vivo. To differentiate between the injected compound and the various metabolic products, an imaging technique capable of separating the different chemical-shift species must be used. In this paper, the design and testing of a pulse sequence for rapid magnetic resonance spectroscopic imaging (MRSI) of hyperpolarized (13)C is presented. The pulse sequence consists of a small-tip excitation followed by a double spin echo using adiabatic refocusing pulses and a "flyback" echo-planar readout gradient. Key elements of the sequence are insensitivity to calibration of the transmit gain, the formation of a spin echo giving high-quality spectral information, and a small effective tip angle that preserves the magnetization for a sufficient duration. Experiments in vivo showed three-dimensional coverage with excellent spectral quality and SNR.     

Wavelet Encoding Spectroscopic Imaging in Small FOV Regime: Comparison to Chemical Shift Imaging at 3?Tesla

We have recently proposed a new magnetic resonance spectroscopic imaging (MRSI) technique called wavelet encoding spectroscopic imaging (WE-SI), and described its implementation on a clinical 1.5?T scanner. This technique is proposed as an alternative to chemical shift imaging (CSI), to decrease acquisition time, and voxel contamination. The proposed method is implemented here on a clinical 3?T scanner. Phantom and in vivo studies are chosen to validate the technique at higher field, as well as to fully explore the usefulness of this technique, and find its niche of application in the chain of existing MRSI techniques. In wavelet encoding, a set of dilated and translated wavelets are used to span a localized space by dividing it into a set of sub-spaces with pre-determined sizes and locations. Due to their simple shapes, Haar wavelets are chosen. They are represented in the modified PRESS sequence by the selective excitation and refocusing radio-frequency (RF) pulses. The wavelets dilation and translation are achieved by changing the strength of the localization gradients and frequency shift of the RF pulses, respectively. Data acquisition time is reduced using the minimum recovery time when successive MR signals from adjacent sub-spaces are collected. The results obtained at 3?T confirm those obtained at 1.5?T, and demonstrate that despite the low signal-to-noise ratio, the proposed WE-SI provides accurate results and reduces both voxel contamination and acquisition time as compared to CSI. This applies especially in the small field-of-view regime where only a small number of voxels is required.     

Multiband excitation pulses for hyperpolarized 13C dynamic chemical-shift imaging

Hyperpolarized 13C offers high signal-to-noise ratios for imaging metabolic activity in vivo, but care must be taken when designing pulse sequences because the magnetization cannot be recovered once it has decayed. It has a short lifetime, on the order of minutes, and gets used up by each RF excitation. In this paper, we present a new dynamic chemical-shift imaging method that uses specialized RF pulses designed to maintain most of the hyperpolarized substrate while providing adequate SNR for the metabolic products. These are multiband, variable flip angle, spectral-spatial RF pulses that use spectral selectivity to minimally excite the injected prepolarized 13C-pyruvate substrate. The metabolic products of lactate and alanine are excited with a larger flip angle to increase SNR. This excitation was followed by an RF amplitude insensitive double spin-echo and an echo-planar flyback spectral-spatial readout gradient. In vivo results in rats and mice are presented showing improvements over constant flip angle RF pulses. The metabolic products are observable for a longer window because the low pyruvate flip angle preserves magnetization, allowing for improved observation of spatially varying metabolic reactions.