Objective: The recent identification of functional depots of brown adipose tissue (BAT) in adult humans has potential implications for the treatment of obesity. In order to evaluate new therapies aimed at inducing the production of more BAT or activating BAT in humans, it will be important to develop noninvasive methods to assess the functional state of the tissue in vivo . In this study, we investigate the feasibility of using hyperpolarized 13 C imaging to noninvasively identify functional, activated BAT in an in vivo rodent model, in less than 1 min, following an infusion of pre-polarized [1- 13 C] pyruvate. Design: Hyperpolarized 13 C imaging was used to monitor BAT metabolic conversion of pre-polarized [1- 13 C] pyruvate in rats during baseline and norepinephrine (NE)-stimulated conditions. Results: Activated BAT, stimulated by NE injection, can be detected in rats by increased conversion of pre-polarized [1- 13 C] pyruvate into its downstream products 13 C bicarbonate and [1- 13 C] lactate. The colocalization of the 13 C signal to interscapular BAT was validated using hematoxylin–eosin histological staining. Conclusion: The radiation-free nature and recent translation into the clinic of the hyperpolarized 13 C-imaging test may potentially facilitate trials of therapeutics targeting BAT activation in humans.

A. P. Chen, R. E. Hurd, M. A. Schroeder, A. Z. Lau, Y-P. Gu, W. W. Lam, and C. H. Cunningham GE Healthcare, Toronto, ON, Canada, GE Healthcare, Menlo Park, CA, United States, Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, United Kingdom, Imaging Research, Sunnybrook Health Sciences Centre, Toronto, ON, Canada, Deptartment of Medical Biophysics, University of Toronto, Toronto, ON, Canada

Introduction Hyperpolarized C substrates have become a promising tool to study real-time metabolic processes in vivo, particularly in the heart. This was first shown using hyperpolarized C pyruvate to characterize cardiac metabolism noninvasively in the pig using a single slice chemical shift imaging (CSI) technique [1]. Recently, rapid multislice imaging of hyperpolarized C pyruvate and bicarbonate was demonstrated by Lau et al. [2], using a single shot spiral pulse sequence. To be able to acquire such images rapidly with a good clinical value it is important to optimize the RF coils to obtain the best signal-to-noise ratio (SNR) possible. Even though the sample losses are known to dominate over the coil losses for frequencies higher than 20 MHz, this strongly depends on the coil/sample size [3]. Surface coils offer considerable gains in SNR compared to whole-body coils, but their performance is limited to the volume nearby and up to a distance of about one radius [3]. This is why in the case of surface coils not only the sample and coil losses are important, but also the homogeneity of the RF field through the volume of interest and the distance to it. The objective of this work was to characterize a transmit/receive surface coil for hyperpolarized C imaging of in vivo cardiac metabolism in the pig. The quality factors (Q) of the coil loaded and unloaded were measured to estimate the sample/coil losses through the loading factor of the coil. Sensitivity maps of the surface coil were computed and compared to a double channel surface coil design. The actual RF map of the single surface coil was measured by imaging a homogenous spherical phantom and was compared to the theoretical maps. Hyperpolarized C pyruvate images of the heart were also obtained to estimate the SNR in vivo. Methods A custom-built transmit/receive C surface coil of 13 cm diameter was used for all the experiments in this work. The quality factors (Q) of the coil in loaded and unloaded conditions were measured using a network analyzer (Agilent Technologies 4395A, USA). The loading measurement was performed by placing the coil over the pig chest in the same position as used for imaging. Three dimensional sensitivity maps of the custom-built coil and a simulated dual-channel surface coil were computed in Matlab (The MathWorks Inc., Massachusetts, USA), by applying Biot-Savart law, and compared. All imaging experiments were performed on a GE MR750 3T MR scanner (GE Healthcare, Waukesha, WI). Real RF/sensitivity maps of the single transmit/receive surface coil were obtained by imaging a homogenous spherical phantom of 10 cm diameter filled with a mixture of deionized water and ~3 mL of 80 mM hyperpolarized C pyruvate. The scan was started 15 s after injection of hyperpolarized C to allow mixing, and the SNR maps were compared to the theoretical sensitivity maps computed before. The same sequence (i.e. 48 cm FOV, 8 mm in-plane resolution, 1 cm slice thickness ,single-shot spiral, 16834 samples over 65.5 ms with 250 kHz sampling [2]) was used for imaging hyperpolarized [1-C] pyruvate in both, phantom and animal studies. In vivo images were obtained in a specific pathogen free (SPF) pig (25 kg) under a protocol approved by the institutional animal care and use committee. The injection was 15 mL of 83 mM pre-polarized [1-C] pyruvate as described in [2]. For anatomical landmarking, cardiac-gated breath-held SSFP CINE images were acquired in the short-axis view (TR = 4.2 ms, TE = 1.8 ms, FOV 24 cm, slice thickness 5 mm, spacing 5 mm, matrix size 224×224) using a separate H surface coil [2]. In vivo hyperpolarized [1-C] pyruvate images of the heart were acquired using the single-shot cardiac-gated spiral pulse sequence described before. Care was taken not to move the pig when the coil was replaced between proton and C acquisitions. The SNR values of C images were estimated in multiple region-of-interests (ROIs) using the mean value of the image signal and divided by the standard deviation of the noise in the background. Results and Discussion Q factors of 45 and 237 were measured on the single transmit/receive C surface coil for loaded and unloaded conditions respectively. This corresponds to a loading factor of 0.81, meaning that in fact the noise is sample dominated and the losses due to the coil are negligible as expected for these frequencies. Figure 1 shows the RF map for the single surface coil (A, B and C) and the two channel surface coil (D, E and F), as simulated in Matlab. The yellow lines represent the RF coils. Due to the planar geometry of the coil the sensitivity decreases with the distance as shown in Fig. 1 (A, B and C), hence the signal drops below the noise floor for the distal regions of the heart. By using a dual coil, with similar dimensions but with positioning along the sides of the pig’s chest (Fig. 1 D, E and F) the homogeneity of the RF field can be dramatically improved through the entire heart, permitting a better excitation of the entire volume (e.g. ~same flip angle) and the improvement of the SNR in the posterior part of the heart between 3 and 4 times as seen in Fig. 1 (please note color scale). Figure 2 shows hyperpolarized [1-C] pyruvate images and SNR distribution as a function of the distance to the coil for the spherical phantom (A and B) and the pig heart in vivo (C and D) using the single transmit/receive surface coil. The experimental results using the homogenous phantom (Fig. 2 A and B) agree well with the simulation results (Fig. 1). The SNR decayed close to zero at a distance of one diameter from the coil as seen in Fig. 2 C and D. A SNR of 44 was measured at the anterior myocardium (i.e. closer to the coil), decaying to 20 by the mid-heart and ~0 by the posterior end as shown in Fig. 2 C and D. Note that the SNR distribution is not only affected by the sensitivity of the coil but also by the effective flip angle throughout the image since the coil is used in transmit/receive mode. However, the SNR measured in the proximal region and the simulation results can be used to predict that a two-coil system would give an SNR of at least 20 throughout the entire heart and possibly higher with a proper flip angle calibration. The best approach would be to use a transmit-only receive-only coil configuration (TORO) using a whole-body transmit coil to ensure the best homogeneity of the RF field and receive the signal using the dual channel coil simulated in this work. Conclusions A single channel C transmit/receive surface coil was characterized with benchtop and in vivo measurements. Even though sample losses dominate in our experiments, further improvement of the SNR throughout the volume of interest can be achieved by using the dual channel surface coil simulated in this work. This coil configuration will allow the imaging of the different metabolite signals (e.g. [1-C] pyruvate, [C] bicarbonate) even in the posterior regions of the myocardium, which is not possible at this moment with the single channel surface coil used in this work. In the near future the dual channel coil simulated in this work will be constructed as well as a full body transmit coil. References: 1Golman et al. Magn Reson Med (59) 2008. 2Lau et al. Magn Reson Med (64) 2010. 3Doty et al. NMR in biomedicine (20) 2007. 4Hayes et al. Med. Phys. (12) 1985.

Introduction Metabolic imaging using 13C-labeled compounds pre-polarized using DNP-dissolution [1,2] requires some form of spectral encoding along with the acquisition of imaging data, which strongly influences the design of pulse sequences for this application. The most straightforward method for accomplishing this spectral and spatial encoding is by conventional chemical-shift imaging. The strategy is flexible in that no prior knowledge about the number and relative frequency separation of the spectral components are needed. However, the underlying pulse sequences are limited in terms of SNR efficiency [3]. Motivated by the need for 3D metabolic images at multiple timepoints, a rapid spectral-spatial echo-planar imaging (ss-epi) pulse sequence tailored for clinical application was developed. Theory In pulse sequences employing spectrally-selective excitation, efficient, long-duration data readouts are required for adequate SNR efficiency. These long readouts are typically sensitive to the unavoidable frequency shifts that occur in practice due to tissue susceptibility differences as well as B0 eddy currents. However, the echo-planar readouts used below behave in a relatively benign fashion, with a spatial shift negligible in all but one imaging direction. The feasibility of using this spatial shift as a method for measuring and correcting the frequencyinduced spatial shifts [4] that occur in vivo was investigated. The idea is that by reversing the k-space trajectory for every other time point, the direction of the spatial shift for a given frequency is reversed. This is demonstrated in Fig. 2. To correct for the resulting mis-registration, the spatial shift that maximally aligns the two shifted images must be found. However, this is non-trivial because the images to be aligned result from different timepoints, and will thus contain different (but similar) pyruvate and lactate distributions. To cope with this, we propose to use mutual information [5] to find the “best” alignment between images. Methods and Results A spectral-spatial excitation pulse was designed with 18.2 ms pulse duration, giving a 120 Hz spectral passband width (FWHM). The RF pulse was implemented with a rapid ‘flyback’ echo-planar readout gradient shown in Fig 1. The image resolution and encoding matrix were tailored to the desired 3D coverage of the human prostate using an endorectral receiver coil [6,7]: 5 mm isotropic spatial resolution, field of view: 8 cm A/P, 8 cm R/L and 6 cm S/I. In vivo studies were performed using a GE MR750 3 T scanner (GE Healthcare, Waukesha, WI) and a micro-strip dual-tuned 1H-13C rat coil (Magvale, San Francisco, CA). All animal experiments followed a protocol approved by the local institutional animal research committee. A HyperSense DNP polarizer (Oxford Instruments, Abingdon, UK) was used to polarize neat (99% purity) [1-13C] pyruvic acid for 1 hour following previously described methods [1]. Tail-vein injections of 2.0mL/80mM of pre-polarized [113C] pyruvate were performed in RNU nude rats (n=3) implanted with U87 tumors in the flank. The injections were 10 seconds, with data acquisition started at the beginning of the injection. Time resolved 3D images of pyruvate, lactate and a urea reference were acquired with 5 s temporal resolution over a 1 minute duration in each study. For each rat, a second injection was performed with the center frequency purposely mis-set by +35 Hz to test the correction for erroneous shifts in the images. Representative images resulting from the pulse sequence are shown in Fig. 3. Note the excellent correspondence between 13C signal and anatomical detail in the T2-weighted images. The spatial shifts required to correct the lactate images at the level of the kidneys, computed by mutual-information registration, are listed in the Table. Even without an intentional frequency error, there are non-zero shifts due to inhomogeneous B0, and these are different for the different rats. Overall, the change in shift induced by the 35 Hz frequency offset was 5.5 +/0.3 mm. This agrees well with the expected 5.7 mm shift based on the 2.02 ms delay between k-space lines (giving 30.9 Hz per pixel). The 0.5 mm range in values corresponds to a frequency sensitivity of 3 Hz. Conclusions A spectral-spatial echo-planar imaging sequence was designed for clinical metabolic imaging and was tested in animal models. A correction for spatial mis-registration due to frequency shifts based on blip-reversal was implemented and validated. Excellent image quality and apparent agreement with the underlying anatomy was observed. The frequency correction method was shown to have an accuracy of 3 Hz.

excite resonances in the C2 pyruvate spectrum. The displayed spectralspatial profiles show the spectral placement of the pulse for (top) C5 glutamate and (bottom) C2 pyruvate. The list of chemical shifts at 3T, relative to C5 glutamate, are (a) 800 Hz for C2 pyruvate, (b) 0 Hz for C5 glutamate, (c) -310 Hz for C2 acetylcarnitine, (d) -410 Hz for C1 pyruvate, (e) -2900 Hz for C2 pyruvate hydrate, and -3600 Hz for C2 lactate. These values were used to design a pulse with 1600 Hz stopband, 200 Hz FWHM passband, pulse length 13.5ms, and 15 mm minimum slice thickness.

Fig. 2. Ordering scheme for simultaneous metabolic imaging and Bloch-Siegert B1 mapping in vivo. Thirty-seven frames are acquired. The shaded boxes show the excited metabolite and nominal flip angle in each frame. The labeled pyruvate frames (*) contain an off-resonance Bloch-Siegert pulse as in Fig. 1. The flip angle is chosen to avoid saturation of signal in future frames and to maximize the SNR of each acquired image. Fig. 3. B1 map acquired on a HP [1-C] pyruvate phantom. The surface coil is at the bottom of the image; a cross-section through the B1 map is shown. The transmit B1 falls off with increasing distance from the coil.

INTRODUCTION: Metabolic imaging using pre-polarized substrates labeled with a 13C nucleus has proven to be a promising new tool (1). Often, chemical-shift imaging (CSI) acquisitions are used to map the 13C resonances over 2D or 3D volumes so that 13C metabolic data from various tissues can be compared. Due to the time constraints imposed by the relatively short lifetime of the hyperpolarized state, the spatial dimensions of these acquisitions are often encoded with small matrix sizes (e.g. 8 x 8 x 16), resulting in a relatively poor point-spread function (PSF). In this abstract, we have explored the use of non-Fourier spatial encoding to improve the PSF in both in-plane dimensions of hyperpolarized 13C CSI acquisitions. THEORY: Non-Fourier spatial encoding makes use of RF pulses as well as gradients to encode spatial information. One advantage of this is the ability to shape the PSF, improving its localization. In this work, we have combined the PSF-Choice technique introduced by Panych et al. (2) with a Hadamard-like technique to encode the two in-plane dimensions, respectively. PSF-Choice uses a 2D RF pulse that is heavily undersampled in one dimension, resulting in replication of the excitation profile across the FOV in the direction of the undersampling. These replicated excitation profiles form the basis functions for spatial encoding one dimension of the FOV. To encode the other in-plane dimension, the fully sampled dimension of the 2D RF pulses can be used to excite multiband excitation profiles in a manner similar to Hadamard encoding. Instead of Hadamard basis functions, discrete Fourier transform (DFT) basis functions (3) were used in the implementation described below. This had the advantage that the usual Fouriertransform reconstruction could be used with the new method as well as the conventionally acquired data below. METHODS: RF and gradient waveforms for encoding a 5.6 cm FOV with a grid of 8 x 8 voxels were designed in Matlab (The Mathworks Inc.). The 64 different RF pulses and 2 (x and y) gradient waveforms were saved in short integer format. An example RF pulse and the two gradient waveforms are shown in Fig. 1. Bloch equation simulations were performed to compute the PSF resulting from these pulses. A double spin-echo CSI pulse sequence (4) was modified to load these files and step through the different pulses during the x and y phase-encoding loops. Experiments were performed on a GE MR750 3T whole-body scanner equipped with the broadband package using a custom-built dual-tuned 13C/1H transmit-receive RF coil (Magvale LLC, San Francisco). To test the quality of the PSF, a phantom experiment was performed using 3 mL syringe filled with 2.7 M [1-13C] glycine. Data were acquired with non-Fourier encoding as well as conventional phase encoding for comparison. In vivo data were acquired from a ~300g Sprague-Dawley rat, with the non-Fourier CSI pulse sequence started 20 s after the start of the injection of 2.0 mL of 80 mM pre-polarized pyruvate through a tail-vein catheter. For this experiment the FOV was 6.4 cm. RESULTS: Simulated excitation profile from one of the 64 pulses designed for the 8x8 non-Fourier spatial encoding and reconstructed PSF from this scheme are shown in Fig.2. Excellent PSF is achieved in both spatial dimension with the proposed method. Phantom 2D CSI data (Fig.3) demonstrated improvement in localization of the non-Fourier spatial encoding as compared to standard phase encoding. CONCLUSIONS: A non-Fourier method for spatial encoding (in two spatial dimension) in CSI acquisitions for prepolarized 13C metabolic studies was developed and demonstrated. Phantom experiments showed an improved pointspread function and a rat study showed the feasibility of using the method for in vivo data acquisition. REFERENCES: 1. Golman K. et al. PNAS 2006;103:11270-11275. 2. Panych LP. et al. MRM 2005;54:159168. 3. Cunningham CH et al. MRM 2001 45:118-127 4. Cunningham CH. et al. 2007, JMR 2007;287:357-62. Figure 1. Example RF pulse and gradient waveform (1 of 64). Similar to the original PSF-Choice implementation, the RF excitation consists of two RF pulses whose relative amplitude is varied as the y phase-encode steps are acquired. However, here the individual RF pulses excite multiple bands with relative phases corresponding to the elements of a DFT encoding matrix.

INTRODUCTION Dynamic Nuclear Polarization (DNP) of metabolically active 13C-labelled substrates has been reported as a method of generating MR images of in vivo cellular metabolism [1] and hence has the potential to characterize aggressive cancers that exhibit heightened energy metabolism non-invasively[2]. In a phenomenon known as the Warbrug effect, aggressively proliferating tumor cells preferentially up-regulate lactate-dehydrogenase (LDH) enzyme to consume glucose at a high rate and release lactate in glycolysis despite the presence of adequate oxygen [3,4]. It has been demonstrated that quantitative pyruvatelactate exchange rates can be obtained by modeling the temporally resolved hyperpolarized dynamic 13C MR spectroscopic data [5] and it may be feasible to create an in vivo map of enzymatic rates if 13C images of the various metabolites can be acquired dynamically over time. In this abstract, we describe the use of a metabolite specific rapid imaging pulse sequence for acquisition of spatially and temporally resolved hyperpolarized pyruvate and lactate images in vitro and in vivo. The dynamic imaging data are fitted to a kinetic two-pool model and compared to results from dynamic MR spectroscopic data.

One of the challenges of optimizing signal‐to‐noise ratio (SNR) and image quality in 13 C metabolic imaging using hyperpolarized 13 C‐pyruvate is associated with the different MR signal time‐courses for pyruvate and its metabolic products, lactate and alanine. The impact of the acquisition time window, variation of flip angles, and order of phase encoding on SNR and image quality were evaluated in mathematical simulations and rat experiments, based on multishot fast chemical shift imaging (CSI) and three‐dimensional echo‐planar spectroscopic imaging (3DEPSI) sequences. The image timing was set to coincide with the peak production of lactate. The strategy of combining variable flip angles and centric phase encoding (cPE) improved image quality while retaining good SNR. In addition, two aspects of EPSI sampling strategies were explored: waveform design (flyback vs. symmetric EPSI) and spectral bandwidth (BW = 500 Hz vs. 267 Hz). Both symmetric EPSI and reduced BW trended toward increased SNR. The imaging strategies reported here can serve as guidance to other multishot spectroscopic imaging protocols for 13 C metabolic imaging applications. Magn Reson Med, 2009. © 2009 Wiley‐Liss, Inc.

INTRODUCTION: The molecular probes used in hyperpolarized C spectroscopy have important advantages over contrast agents currently in clinical use, expected to have little or no toxicity in humans, even at relatively high concentrations. This feature is particularly appealing given recent concerns about contrast nephropathy associated with iodinated CT contrast, as well as nephrogenic systemic fibrosis (NSF) seen in patients receiving gadolinium chelates for MR[1,2]. Many new C agents may be appropriate for metabolic imaging in humans. In addition to C pyruvate, C lactate itself is a promising primary agent for cancer imaging[3]. Other recent work has demonstrated in vivo pH mapping using C bicarbonate, by hyperpolarization of C cesium bicarbonate followed by an ion exchange method to exchange most of the Cs for sodium[4]. In this abstract, a method for direct polarization of C sodium bicarbonate is reported, that is suitable for use in humans. This method has been combined with a copolarization approach that allows simultaneous polarization of C bicarbonate and C pyruvate, to perform both pH and metabolic mapping in vivo using a single contrast bolus.

Introduction: Hyperpolarization of spins via DNP has been explored as a method to noninvasively study real-time metabolic processes in vivo using C labeled substrates [1]. Hyperpolarized C has recently been used to image cardiac metabolism non-invasively in vivo in rat [2] and pig [3]. In this abstract, we investigate the use of hyperpolarized C MR to study real-time cardiac metabolism in a pig model of cardiac disease.

Introduction: Development of techniques to retain highly polarized spins in solution via DNP has enabled the use of C labeled metabolic intermediates such as pyruvate to investigate enzymatic exchange processes in vivo with high temporal/spatial resolution (12). To achieve high polarization enhancement in reasonable amount of time (~1hr), the technique thus far has been limited to substrates with low molecular weight and high solubility in organic solvent/water. In conditions where DNP is dominated by solid state and thermal mixing effects, the electron and nuclear interaction is anisotropic and there is a distribution in the DNP enhancement factor across the sample. These local spin temperature differences are then equilibrated via nuclear dipolar interaction (spin diffusion) (3). Thus it may be possible to enhance the polarization of a target substrate that has poor DNP properties (low solubility, poor enhancement) with a secondary substrate that demonstrates high DNP enhancement and can be added in high concentration at the given condition, allowing the more concentrated and highly polarized nuclei to assist the polarization of the target nuclei via this mechanism. The goal of this study was to develop and test such a method.

INTRODUCTION Dynamic Nuclear Polarization (DNP) of metabolically active 13C-labelled substances has been reported as a method of generating MR images of in vivo cellular metabolism [1] and hence has the potential to detect aggressive cancers that exhibit heightened metabolism [2]. By injecting active, hyperpolarized 13C metabolites such as pyruvate, it is possible to visualize their transformation to lactate and other products that have distinctive MRI frequencies. However, in addition to the metabolites that are transformed locally, there are also metabolic products generated in blood and other tissues that enter the tissue of interest via the systemic circulation, complicating analysis of local metabolism. Using hyperpolarized [1-13C] pyruvate and MR spectroscopy, this study attempts to quantify pyruvate-lactate conversion in ex vivo whole blood which might be useful in providing an estimation of blood metabolic contribution to regions of interest in in vivo studies. METHODS Five milliliters of whole blood samples were taken from live Sprague-Dawley rats (n=2) and a pig and stored in lithium heparin coated blood collection tubes (BD vacutainer) immediately before the experiments. Samples of [1-13C] pyruvic acid and 15 mM trityl radical were hyperpolarized at 1.4K with a DNP hyperpolarizer (Oxford Instruments, Tubney Woods, UK). The sample was rapidly dissolved with NaOH/Tris/EDTA buffer solution to a concentration of 80mM, and a bolus of 1.9mL was injected into the blood inside the collection tube at the MRI scanner. Prior to the injection, the vacuum in the collection tube was restored with a 5mL syringe to facilitate the injection. This procedure was tested with dye and water to ensure adequate mixing. The injection took 2-3 seconds while data acquisition began immediately. The experiments were performed on a GE Excite 3T MRI scanner with a dual-tuned transmit-receive rat birdcage coil. An adiabatic double spin echo pulse sequence (TR=1s, TE=35ms, flip angle=10deg) [3] was used in this study. Data analysis was performed in SAGE (GE Healthcare, Waukesha, WI) and Matlab (The Mathworks, Inc., Natick, MA) RESULTS [1-13C] pyruvic acid was transferred from the injected hyperpolarized pyruvate to lactate in the whole blood as expected. The zmagnetization peak integrals can be fit to the modified Bloch equations for a two compartmental linear kinetic model (figure1 ): dP dt = −(kLP + k0P )P + RA, dL dt = kLPP − k0LL Where P , L denote the peak integrals of z-magnetizations of pyruvate and lactate, respectively; t is time; RA is the constant rate of appearance, The peak integrals were SNR weighted, least squares fitted to these equations to obtain RA , the rate constants kLP , and the apparent spin lattice relaxation rates, k0P , and k0L .The metabolites’ spectra and the two compartmental model fit of the metabolic kinetics are shown in figures 2 and 3, respectively. The time to peak for pyruvate and lactate are 3±1s, and 16 ±3s, respectively. Knowing the amount of [1-13C]pyruvate injected and using the parameters from the model, pyruvate-lactate substrate flux can be estimated (see Table 1).

We present for the first time dynamic spectra and spectroscopic images acquired in normal rats at 3T following the injection of (13)C-1-pyruvate that was hyperpolarized by the dynamic nuclear polarization (DNP) method. Spectroscopic sampling was optimized for signal-to-noise ratio (SNR) and for spectral resolution of (13)C-1-pyruvate and its metabolic products (13)C-1-alanine, (13)C-1-lactate, and (13)C-bicarbonate. Dynamic spectra in rats were collected with a temporal resolution of 3 s from a 90-mm axial slab using a dual (1)H-(13)C quadrature birdcage coil to observe the combined effects of metabolism, flow, and T(1) relaxation. In separate experiments, spectroscopic imaging data were obtained during a 17-s acquisition of a 20-mm axial slice centered on the rat kidney region to provide information on the spatial distribution of the metabolites. Conversion of pyruvate to lactate, alanine, and bicarbonate occurred within a minute of injection. Alanine was observed primarily in skeletal muscle and liver, while pyruvate, lactate, and bicarbonate concentrations were relatively high in the vasculature and kidneys. In contrast to earlier work at 1.5 T, bicarbonate was routinely observed in skeletal muscle as well as the kidney and vasculature.

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 cm3 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.