Hyperpolarized [2-13C]pyruvate MR molecular imaging with whole brain coverage

Abstract

Hyperpolarized (HP) ¹³C Magnetic Resonance Imaging (MRI) was applied for the first time to image and quantify the uptake and metabolism of [2-¹³C]pyruvate in the human brain to provide new metabolic information on cerebral energy metabolism. HP [2-¹³C]pyruvate was injected intravenously and imaged in 5 healthy human volunteer exams with whole brain coverage in a 1-minute acquisition using a specialized spectral-spatial multi-slice echoplanar imaging (EPI) pulse sequence to acquire ¹³C-labeled volumetric and dynamic images of [2-¹³C]pyruvate and downstream metabolites [5-¹³C]glutamate and [2-¹³C]lactate. Metabolic ratios and apparent conversion rates of pyruvate-to-lactate (k_PL) and pyruvate-to-glutamate (k_PG) were quantified to investigate simultaneously glycolytic and oxidative metabolism in a single injection.

Keywords: Brain Metabolism; Hyperpolarized carbon-13; Molecular imaging.

Fig. 1.

Singleband spectral-spatial RF pulse and EPI readout showing simulated responses from [2–¹³C]pyruvate, [5–¹³C]glutamate and [2–¹³C]lactate to demonstrate the effectiveness of designed pass & stopbands. A. Chemical shift frequencies of [2–¹³C]pyruvate and downstream metabolites were obtained from an NMR spectrum acquired in a previous study (Chung et al., 2019). The asymmetry of the lactate doublet is primarily due to the RF excitation band used. B. Sequence diagram illustrating the metabolite-selective EPI acquisition. C. The spectral-spatial response of the 2D RF pulse used to selectively excite [2–¹³C]pyruvate (206.00 ppm), [5–¹³C]glutamate (182.29 ppm), and the [2–¹³C]lactate doublet (72.39 ppm and 67.90 ppm) with minimal off-resonance excitation. The vertical dashed lines correspond to the relative frequency for each metabolite at 3T. See Supplementary Fig. 1 for frequency responses of the spectral-spatial RF pulse when centered on other resonances.

Fig. 2.

4D dynamics of HP [2–¹³C]pyruvate, [5–¹³C]glutamate, [2–¹³C]lactate (downfield peak) and [2–¹³C]lactate (upfield peak) from the human brain of a healthy volunteer (Vol-1). Displayed images show the first 10 timeframes with 3 second temporal resolution for a total of window of 30 s following denoising using a patch-based HOSVD method (Kim et al., 2021). Shown on the left are ¹H IR-SPGR anatomy images capturing average of slices. Upper window levels for [2–¹³C]pyruvate data were adjusted to 20% of the maximum intensity. The images were denoised and zero-filled two-fold for display.

Fig. 3.

AUC (Area-Under-Curve) images from the volunteer (Vol-1) summed over 20 timeframes overlaid on ¹H IR-SPGR images. [2–¹³C]lactate images illustrate acquired signal after summing both downfield and upfield peaks. High resolution (0.75 × 0.75 cm²) pyruvate data shows strong arterial and venous signal, and a brain mask to reduce the pyruvate signal from adjacent muscle was consequently applied. The images were denoised and zero-filled for display.

Fig. 4.

Kinetic analyses of dynamic HP [2–¹³C]pyruvate data. (Top row) A sagittal ¹H FLAIR image from another volunteer showing a mid-slice position for ¹³C EPI data denoted with yellow lines (dashed/solid: center/edge of the slice) and denoised ¹³C AUC images for pyruvate, glutamate, and lactate overlaid on a corresponding axial ¹H SPGR image. (Middle row) Signal intensities of pyruvate, glutamate, and lactate over time from the selected voxels indicated in the AUC images. Kinetic fits for lactate and glutamate are displayed. (Bottom row) Calculated k_PL and k_PG maps overlaid on axial ¹H SPGR image were obtained from healthy volunteer brain data. The selected voxels are also indicated in the k_PL and k_PG maps.