Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Nov 1.
Published in final edited form as: Magn Reson Med. 2022 Jun 26;88(5):2190–2197. doi: 10.1002/mrm.29354

Kinetic Analysis of Multi-Resolution Hyperpolarized 13C Human Brain MRI to Study Cerebral Metabolism

Jasmine Y Hu 1,2, Yaewon Kim 1, Adam W Autry 1, Mary M Frost 1, Robert A Bok 1, Javier E Villanueva-Meyer 1, Duan Xu 1,2, Yan Li 1,2, Peder EZ Larson 1,2, Daniel B Vigneron 1,2, Jeremy W Gordon 1
PMCID: PMC9420752  NIHMSID: NIHMS1810382  PMID: 35754148

Abstract

Purpose:

To investigate multi-resolution hyperpolarized (HP) 13C pyruvate MRI for measuring kinetic conversion rates in the human brain.

Methods:

HP [1-13C]pyruvate MRI was acquired in six subjects with a multi-resolution EPI sequence at 7.5 × 7.5 mm2 resolution for pyruvate and 15 × 15 mm2 resolution for lactate and bicarbonate. With the same lactate data, two quantitative maps of pyruvate-to-lactate conversion (kPL) maps were generated: one using 7.5 × 7.5 mm2 resolution pyruvate data and the other using synthetic 15 × 15 mm2 resolution pyruvate data to simulate a standard constant resolution acquisition. To examine local kPL values, four voxels were manually selected in each study representing brain tissue near arteries, brain tissue near veins, white matter, and gray matter.

Results:

High resolution 7.5 × 7.5 mm2 pyruvate images increased the spatial delineation of brain structures and decreased partial volume effects compared to coarser resolution 15 × 15 mm2 pyruvate images. Voxels near arteries, veins and in white matter exhibited higher calculated kPL for multi-resolution images.

Conclusion:

Acquiring HP 13C pyruvate metabolic data with a multi-resolution approach minimized partial volume effects from vascular pyruvate signals while maintaining the SNR of downstream metabolites. Higher resolution pyruvate images for kinetic fitting resulted in increased kinetic rate values, particularly around the superior sagittal sinus and cerebral arteries, by reducing extracellular pyruvate signal contributions from adjacent blood vessels. This HP 13C study showed that acquiring pyruvate with finer resolution improved the quantification of kinetic rates throughout the human brain.

Keywords: Hyperpolarization, MRI, Kinetics, Carbon-13, Pyruvate

Introduction

Hyperpolarized (HP) carbon-13 MR has been investigated in animals since 2006 to provide a unique window into cellular metabolism, enabling the quantification of enzyme-catalyzed conversion rates that inform on critical cellular biochemistry in both normal and pathologic conditions1–3. A first-in-human proof of concept clinical trial of HP [1-13C]pyruvate completed in 2013 demonstrated feasibility and safety in patients4. The subsequent development of commercial research polarizers enabled new technical developments and initial human studies over the past 5 years in a variety of applications including prostate cancer, brain tumors, renal cancer, cardiac disease, pancreatic cancer, traumatic brain injury and breast cancer5–14.

Since 2018, numerous studies have focused on investigating cerebral energy metabolism in the normal human brain and neuro-pathologies, demonstrating novel insights into brain bioenergetics by measuring HP pyruvate conversions to lactate and bicarbonate6,10,11,15–17. Conversion from pyruvate to lactate and to bicarbonate provides a measure of metabolic preference for either oxidative phosphorylation, which generates bicarbonate, or glycolytic metabolism, which generates lactate. The normal brain produces both lactate and bicarbonate, and increased lactate and decreased bicarbonate production were observed in brain tumor lesions6,15. These studies support the utility of HP pyruvate metabolism as a biomarker of brain tumor metabolic reprogramming and response to therapy.

Recent technical advances for HP studies have shown that metabolic quantification could be further improved through specialized acquisition and denoising techniques18–21. These techniques provide improved SNR and finer spatial resolution, enabling better localization of metabolism. In particular, a multi-resolution HP 13C EPI approach was developed that enables the acquisition of the injected HP 13C pyruvate at a higher spatial resolution than its metabolic products lactate and bicarbonate with their lower inherent SNR18. The purpose of this study was to apply for the first time this multi-resolution approach and specialized analysis to investigate HP pyruvate-to-lactate conversion rates, kPL, in the normal brain with the motivation that higher pyruvate resolution could reduce partial volume effects of extracellular HP pyruvate in cerebral blood vessels that could cause errors in intracellular kPL measurements.

Methods

HP brain studies were performed in six healthy human volunteer subjects with informed consent according to University of California San Francisco IRB and FDA IND approved protocols. The mean age of the subjects was 41.2 years (range 29–61 years) and all subjects were male. Before each study, a pharmacist prepared samples containing 1.47 g of Good Manufacturing Practice grade [1-13C]pyruvic acid (MilliporeSigma Isotec) and 15 mM trityl electron paramagnetic agent (EPA; AH111501, GE Healthcare). The samples were polarized using a SPINlab polarizer (GE Healthcare) operating at 5 T and 0.8 K for > 2 hours. After dissolution of the polarized samples, the EPA was removed by filtration, the solution was neutralized with a Tris-buffered NaOH solution, and the quality control parameters of pH, pyruvate and residual EPA concentrations, polarization, and temperature were measured prior to injection. In parallel, the sterile filter (0.2 μm, ZenPure, Manassas, VA) was tested in agreement with manufacturer specifications prior to injection. After release by a pharmacist, a 0.43 mL/kg dose of ~250 mM pyruvate was injected at a rate of 5 mL/s, followed by a 20 mL sterile saline flush (0.9% sodium chloride, Baxter Healthcare Corporation).

Studies were performed on a 3T MR scanner (MR750, GE Healthcare) using an integrated 8 channel 1H/24 channel 13C phased array receiver with an 8-rung low-pass 13C volume transmit coil (Rapid Biomedical, Würzburg, Germany). Hyperpolarized 13C data were acquired with a metabolite-selective imaging approach, using a singleband spectral-spatial RF pulse for excitation (passband FWHM = 130 Hz, stopband = 868 Hz) and a single-shot symmetric echoplanar readout for encoding22. Scan parameters were 125 ms TR, 30.7 ms TE, 32 × 32 matrix size, ±19.23 kHz BW, 1.064 ms echo-spacing, and eight slices with an axial orientation. Data acquisition started 5 seconds after the end of the saline injection for the first three subjects and immediately after the end of the saline injection for the latter three subjects. Pyruvate was excited with a 20° flip angle and lactate and bicarbonate were excited with a 30° flip angle. The slice thickness was 20 mm for the first subject and 15 mm for the other five subjects. The in-plane spatial resolution for each metabolite was changed by independently scaling the encoding gradients, resulting in 7.5 × 7.5 mm2 resolution for pyruvate and 15 × 15 mm2 resolution for lactate and bicarbonate. Twenty time points were acquired with a 3 second temporal resolution for a total scan time of one minute. Immediately following imaging, a non-localized spectrum was acquired to confirm the center frequency was set correctly. For anatomic reference, 1H 3D inversion-recovery spoiled gradient-recalled echo (IR-SPGR) was acquired with the dual-tuned coil. 1H 3D IR-SPGR scan parameters were 6.7 ms TR, 2.5 ms TE, 450 ms IR time, 25.6 × 25.6 × 18.6 cm2 FOV, 256 × 256 × 124 matrix size (1 × 1 × 1.5 mm3 resolution).

The 13C EPI data were reconstructed using the Orchestra toolbox (GE Healthcare). Multichannel data were pre-whitened23 and then coil combined using pyruvate to estimate the coil weights24. Denoising was performed on the coil-combined data using global-local higher-order singular value decomposition (GL-HOSVD) as described in Kim et al. for hyperpolarized MRI20. To quantify the reduction in partial volume effects with variable-resolution EPI, a synthetic 15 × 15 mm2 pyruvate data set was obtained by cropping the central 16 × 16 region of k-space, zeropadding to the original matrix size (32 × 32), Fermi filtering to minimize Gibbs ringing, and transforming back to the image domain. Lactate and bicarbonate images were cropped and zeropadded to match the pyruvate FOV and matrix size, and signal values were normalized to voxel volume to account for the different acquisition resolution. Proton images were used in the FSL FAST algorithm25 to generate brain masks, and the proton images were also summed in the slice dimension to match the carbon slice thickness.

Kinetic rate constants for each voxel were computed using an inputless two-site model to generate quantitative maps of pyruvate-to-lactate conversion (kPL)26. With the same lactate data, multi-resolution and constant-resolution kPL maps were generated: one using the 7.5 × 7.5 mm2 resolution pyruvate data and the other using the synthetic 15 × 15 mm2 resolution pyruvate data. Kinetic rate maps were thresholded to select voxels that had a lactate area-under-curve (AUC) SNR > 5 and fitting error < 30%. Here we used the fitting error as defined in Mammoli et al26, where an absolute fitting error is defined to be half of the 95% confidence interval range from the nonlinear least squares fitting and the fitting error is the absolute error divided by the fitted rate constant. Due to limited kPB coverage in white matter after SNR and fitting error thresholding, kPB maps were not considered for this analysis. kPL percent differences between the different resolutions were calculated by taking the difference of multi-resolution kPL and constant-resolution kPL and dividing by the constant-resolution kPL on a voxel-wise basis. Lactate-to-pyruvate AUC ratios were also calculated using both pyruvate resolutions. To examine local kinetic rates across the six subjects for both resolution schemes, four voxels were manually selected in each study representing brain near arteries, brain near veins, white matter and gray matter. Statistical analysis with two-sided Wilcoxon signed rank tests was used to compare the paired multi-resolution and constant-resolution voxels.

Results

High resolution 7.5 × 7.5 mm2 pyruvate images exhibit increased spatial delineation of brain structures and decreased partial volume effects compared to the coarse resolution 15 × 15 mm2 pyruvate images as shown in Figure 1. In particular, high pyruvate signals from the gray matter, cerebral arteries, and sagittal sinus spill over into surrounding brain areas in the coarse resolution images. The lactate and bicarbonate images exhibit adequate SNR > 5 at 15 × 15 mm2 resolution.

Figure 1.

Figure 1.

Open in a new tab

Hyperpolarized 13C pyruvate, lactate and bicarbonate signals summed over 60 seconds with reference proton images of subject 2. The 15 × 15 mm2 pyruvate data set was obtained by cropping 7.5 × 7.5 mm2 pyruvate k-space data, zerofilling, and applying a Fermi filter. The downsampled pyruvate images exhibited partial volume spillover of signal particularly in the vasculature, including the middle cerebral arteries indicated by red arrows in slices 3–4 and the superior sagittal sinus indicated by blue arrows in slices 3–7. Carbon images were zerofilled once for display and intensity units are arbitrary.

In this study, dynamic high resolution pyruvate images also improve identification of pyruvate arrival in cerebral blood vessels. In Figure 2 the pyruvate bolus is seen first in the internal carotid and middle cerebral arteries, followed by the transverse and sagittal sinuses after one 3-second time frame. The arterial pyruvate signal fades after 15 seconds, whereas the venous pyruvate signal is maintained for the duration of the 60-second acquisition.

Figure 2.

Figure 2.

Open in a new tab

Hyperpolarized 13C pyruvate dynamic images of pyruvate delivery in the brain with reference proton images of subject 5. The colored arrows point out arrival of pyruvate in the anterior cerebral circulation (internal carotid, red; middle cerebral arteries, yellow) and transit into the superior sagittal (blue) and transverse dural venous sinuses (green). The arterial pyruvate signal disappears approximately 15 seconds after arrival, while the venous pyruvate signal remains high and visible towards the end of the acquisition. Carbon images were zerofilled once for display and intensity units are arbitrary.

The increased spatial delineation of high resolution pyruvate images is also apparent in the multi-resolution kPL maps in Figure 3. The kPL values were higher for multi-resolution kinetic maps, especially for the pons in the most inferior slice and white matter in the 3rd and 4th slices. The partial volume effect of decreased kPL was most apparent near the ventricles and sagittal sinus, regions where high pyruvate signal spilled over from highly perfused gray matter and the sagittal sinus. Supporting Information Table S1 summarizes the mean and standard deviation (SD) of kPL and lactate-to-pyruvate AUC ratios in the brain for each subject. Multi-resolution kPL and AUC ratio means and SDs were significantly higher (p < 0.05) than constant-resolution kPL and AUC ratio means and SDs. Average differences for multi-resolution as compared to constant-resolution were: 19% and 70% increase in kPL mean and SD; 22% and 93% increase in AUC ratio mean and SD. For kPL calculated from signals summed over the whole brain, the average percent kPL difference across subjects was 3% higher for multi-resolution data than for constant-resolution data.

Figure 3.

Figure 3.

Open in a new tab

Kinetic rate maps for pyruvate-to-lactate conversion (kPL, s−1) in subject 1, using 15 × 15 mm2 lactate signals with 7.5 × 7.5 mm2 pyruvate (multi-resolution) and synthetic 15 × 15 mm2 pyruvate (constant-resolution) brain-masked images, along with reference proton images. The constant-resolution kPL map showed smoothing and loss of fine detail of the kinetic rates in the brain as compared to the multi-resolution kPL map.

To compare kPL between multi-resolution and constant-resolution datasets on a regional basis, we selected voxels in each subject near the arteries and veins, in addition to voxels in white and gray matter. Supporting Information Table S2 summarizes the kPL values in the selected voxels for all subjects. Figure 4 shows the selected voxel positions in proton images and kPL maps for one subject. In this particular volunteer, multi-resolution kPL values were 169, 65 and 44 percent higher than constant-resolution kPL values for voxels near arteries, near veins and in white matter respectively. These kPL differences are also present for the selected voxels in all subjects (p < 0.05), summarized in Table 1. The same effect of the multi-resolution approach is also observed for lactate-to-pyruvate AUC ratios, reported in Supporting Information S3. This difference in kPL values and AUC ratios indicates that less extracellular pyruvate signal was present in the higher-resolution pyruvate voxels than in the coarse-resolution pyruvate voxels, since the lactate signal used for the multi-resolution and constant-resolution kPL maps was identical.

Figure 4.

Figure 4.

Open in a new tab

Kinetic rate maps for pyruvate-to-lactate conversion (kPL, s−1) for subject 5 with four selected voxel positions shown in multi-resolution and constant-resolution maps. The voxels numbered 1–4 refer to locations near artery, near vein, in white matter, and in gray matter respectively. Voxels near arteries, veins and in white matter exhibited 169, 65, and 44 percent higher kPL for multi-resolution maps respectively, indicating less pyruvate signal contributions from blood vessels. The voxel in gray matter showed 7 percent kPL decrease for multi-resolution, indicating less change in the regional pyruvate signal between higher and coarser resolutions.

Table 1.

Summary of percent differences in pyruvate-to-lactate conversion (kPL) between multi-resolution and constant-resolution data from six subjects. Single voxels near arteries, veins, in white matter, and in gray matter were selected manually. Multi-resolution kPL was significantly higher for voxels near arteries and veins and voxels in white matter (p < 0.05), showing a decrease in partial volume spillover of vascular pyruvate. There were inconsistent differences for gray matter voxels. Subject 4 values were an average of two hyperpolarized C13 pyruvate injections in the same exam and stars indicate statistical significance (p < 0.05).

Subject Near Artery
% kPL Difference		 Near Vein
% kPL Difference		 White Matter
% kPL Difference		 Gray Matter
% kPL Difference
1 52 77 47 1
2 57 96 75 46
3 37 106 57 51
4 32 73 27 −16
5 169 65 44 −7
6 153 30 71 −12
Mean ± SD 83 ± 61* 75 ± 27* 54 ± 18* 10 ± 30
Open in a new tab

Discussion

In this study, we utilized a multi-resolution acquisition to determine if higher pyruvate resolution could reduce the inclusion of extracellular pyruvate signals in blood vessels that is not metabolized like intracellular pyruvate in the brain and thus can cause errors in kPL quantification. Higher spatial resolution for pyruvate (7.5 × 7.5 mm2 versus 15 × 15 mm2) decreased the partial volume spillover of vascular pyruvate signals into nearby brain voxels, and as a result the kPL and lactate-to-pyruvate AUC ratios of those voxels calculated from multi-resolution data was higher than those calculated from constant-resolution data. Although the average kPL difference between resolutions for the whole brain was only 3%, for voxels near cerebral arteries the average kPL difference was up to 83%. This indicates that the partial volume effect of coarser resolution changes pyruvate signal magnitudes in a localized manner. Because of this localization, the magnitude of difference between resolutions is small when looking at the whole brain versus examining individual voxels. This partial volume effect was most apparent for voxels in white matter and voxels near cerebral arteries and veins. Using multi-resolution images for kinetic quantification provided more accurate kPL and lactate-to-pyruvate AUC ratio estimates, especially near cerebral blood vessels with large amounts of pyruvate that are not available for metabolic conversion.

In addition, the multi-resolution acquisition scheme takes advantage of the high pyruvate SNR inherent to HP pyruvate studies by acquiring the injected substrate at higher resolution. These higher resolution pyruvate images improve the visualization of neuro-vascular structures and could potentially be used to extract perfusion-weighted information in addition to metabolic conversion27. The GL-HOSVD denoising also improves SNR for all metabolites20, potentially enabling even further improvements in spatial resolution for future studies. Combining a multi-resolution acquisition with spatiotemporal denoising provides both high SNR and high spatial resolution, which in turn improves the accuracy of metabolic quantification.

In relation to previous HP studies on normal brain metabolism, this work studies the impact of multi-resolution acquisition on kPL values and lactate-to-pyruvate AUC ratios. Prior studies have reported on the regional distribution of HP lactate production in terms of lactate-to-pyruvate AUC ratio, lactate z-score, and kPL16,17. While there are multiple metrics that can be used to quantify HP data, using multi-resolution images improves kPL and lactate-to-pyruvate AUC ratio metabolic quantification due to reduced partial volume effects. The multi-resolution EPI approach is also flexible, as the resolutions and flip angles can be customized to each metabolite depending on SNR. Other metabolite-selective methods used for HP imaging such as spiral and balanced steady state free precession offer benefits in terms of speed and SNR and are also compatible with a multi-resolution imaging scheme.

Acquiring pyruvate with finer resolution allows for improved visualization of neurovascular structures and more accurate quantification of metabolite kinetics, but the multi-resolution acquisition comes with certain drawbacks. For this study, the metabolite signals were normalized to their respective voxel volumes for quantification. This method of matching the different resolutions assumes uniform signal distribution throughout voxels, which is not true for every voxel. Increasing the pyruvate resolution with a single-shot readout entails a longer readout time, which can lead to increased geometric distortion in regions of high B0 inhomogeneity. This could be ameliorated by incorporating a field map into the reconstruction or alternating the blip direction during the dynamic acquisition28,29. Limitations of this study also include the lack of in vivo constant-resolution data to directly compare to the acquired multi-resolution data. Previous studies with constant-resolution echoplanar and spiral trajectories used different echo times, which may affect the calculated kinetic rates due to differences in T2*30 and require a correction to directly compare with the values from this study. Despite the variation in acquisition delays employed across this study, the kinetic modeling of kPL used did not require estimation of an input function and was therefore less sensitive to bolus arrival time, so long as the inflow of pyruvate was sufficiently captured31. The 0 second and 5 second delays both enabled sufficient pyruvate signal to be captured for the purposes of fitting.

Conclusions

Acquiring hyperpolarized 13C pyruvate studies with a multi-resolution approach minimizes partial volume effects from vascular pyruvate signals while maintaining the SNR of downstream metabolites. Utilizing higher resolution pyruvate images for kinetic fitting reduced partial volume effects and increased the calculated kinetic rate values, particularly around the superior sagittal sinus and cerebral arteries where high pyruvate signals spill over from the large blood vessels. This hyperpolarized 13C data showed that acquiring pyruvate with finer resolution improved the quantification of kinetic rates throughout the human brain.

Supplementary Material

supinfo

Supporting Information Table S1. Mean and standard deviation (SD) of pyruvate-to-lactate conversion (kPL) and lactate-to-pyruvate area-under-curve (AUC) ratios for multi-resolution and constant-resolution C13 pyruvate data in the brain for six subjects. Multi-resolution kPL and AUC ratio means and SDs were significantly higher (p < 0.05) than constant-resolution kPL and AUC ratio means and SDs. Average differences for multi-resolution as compared to constant-resolution were: 19% and 70% increase in kPL mean and SD; 22% and 93% increase in AUC ratio mean and SD.

Supporting Information Table S2. Pyruvate-to-lactate conversion (kPL) values for multi-resolution and constant-resolution C13 pyruvate data from four regions in the brain. Single voxels near arteries, veins, in white matter, and in gray matter were selected manually. Multi-resolution kPL was significantly higher for voxels near arteries and veins and voxels in white matter (p < 0.05), showing a decrease in partial volume spillover of vascular pyruvate. There were inconsistent differences for gray matter voxels. Subject 4 values were an average of two C13 pyruvate injections in the same exam and stars indicate statistical significance (p < 0.05).

Supporting Information Table S3. Lactate-to-pyruvate area-under-curve (AUC) ratios for multi-resolution and constant-resolution C13 pyruvate data from four regions in the brain. Single voxels near arteries, veins, in white matter, and in gray matter were selected manually. Multi-resolution AUC ratios were significantly higher for voxels near arteries and veins and voxels in white matter (p < 0.05), showing a decrease in partial volume spillover of vascular pyruvate. There were inconsistent differences for gray matter voxels. Subject 4 values were an average of two C13 pyruvate injections in the same exam and stars indicate statistical significance (p < 0.05).

Acknowledgements

This research was supported by NIH grants (PO1-CA118816, U01-EB026412, P41-EB013598) and the UCSF NICO project. Special thanks to the Aarhus University Magnetic Resonance Research Center for contributing to the data collection process.

References

  • 1.Golman K, in ‘t Zandt R, Thaning M. Real-time metabolic imaging. Proceedings of the National Academy of Sciences. 2006;103(30):11270–11275. doi: 10.1073/pnas.0601319103 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Kurhanewicz J, Vigneron DB, Ardenkjaer-Larsen JH, et al. Hyperpolarized 13C MRI: Path to Clinical Translation in Oncology. Neoplasia. 2019;21(1):1–16. doi: 10.1016/j.neo.2018.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Wang ZJ, Ohliger MA, Larson PEZ, et al. Hyperpolarized 13 C MRI: State of the Art and Future Directions. Radiology. 2019;291(2):273–284. doi: 10.1148/radiol.2019182391 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Nelson SJ, Kurhanewicz J, Vigneron DB, et al. Metabolic Imaging of Patients with Prostate Cancer Using Hyperpolarized [1–13C]Pyruvate. Science Translational Medicine. 2013;5(198):198ra108–198ra108. doi: 10.1126/scitranslmed.3006070 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Apps A, Lau JYC, Miller JJJJ, et al. Proof-of-Principle Demonstration of Direct Metabolic Imaging Following Myocardial Infarction Using Hyperpolarized 13C CMR. JACC: Cardiovascular Imaging. 2021;14(6):1285–1288. doi: 10.1016/j.jcmg.2020.12.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Autry AW, Gordon JW, Chen HY, et al. Characterization of serial hyperpolarized 13C metabolic imaging in patients with glioma. NeuroImage: Clinical. 2020;27:102323. doi: 10.1016/j.nicl.2020.102323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Crane JC, Gordon JW, Chen H, et al. Hyperpolarized 13 C MRI data acquisition and analysis in prostate and brain at University of California, San Francisco. NMR in Biomedicine. 2021;34(5). doi: 10.1002/nbm.4280 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Gallagher FA, Woitek R, McLean MA, et al. Imaging breast cancer using hyperpolarized carbon-13 MRI. Proc Natl Acad Sci USA. 2020;117(4):2092–2098. doi: 10.1073/pnas.1913841117 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Granlund KL, Tee SS, Vargas HA, et al. Hyperpolarized MRI of Human Prostate Cancer Reveals Increased Lactate with Tumor Grade Driven by Monocarboxylate Transporter 1. Cell Metabolism. 2020;31(1):105–114.e3. doi: 10.1016/j.cmet.2019.08.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hackett EP, Pinho MC, Harrison CE, et al. Imaging Acute Metabolic Changes in Patients with Mild Traumatic Brain Injury Using Hyperpolarized [1–13C]Pyruvate. iScience. 2020;23(12):101885. doi: 10.1016/j.isci.2020.101885 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Miloushev VZ, Granlund KL, Boltyanskiy R, et al. Metabolic Imaging of the Human Brain with Hyperpolarized 13 C Pyruvate Demonstrates 13 C Lactate Production in Brain Tumor Patients. Cancer Res. 2018;78(14):3755–3760. doi: 10.1158/0008-5472.CAN-18-0221 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Stødkilde-Jørgensen H, Laustsen C, Hansen ESS, et al. Pilot Study Experiences With Hyperpolarized [1- 13 C]pyruvate MRI in Pancreatic Cancer Patients. J Magn Reson Imaging. 2020;51(3):961–963. doi: 10.1002/jmri.26888 [DOI] [PubMed] [Google Scholar]
  • 13.Tang S, Meng MV, Slater JB, et al. Metabolic imaging with hyperpolarized 13 C pyruvate magnetic resonance imaging in patients with renal tumors—Initial experience. Cancer. 2021;127(15):2693–2704. doi: 10.1002/cncr.33554 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Tran M, Latifoltojar A, Neves JB, et al. First-in-human in vivo non-invasive assessment of intra-tumoral metabolic heterogeneity in renal cell carcinoma. BJR|case reports. 2019;5(3):20190003. doi: 10.1259/bjrcr.20190003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Park I, Larson PEZ, Gordon JW, et al. Development of methods and feasibility of using hyperpolarized carbon-13 imaging data for evaluating brain metabolism in patient studies: Hyperpolarized Carbon-13 Metabolic Imaging of Patients With Brain Tumors. Magn Reson Med. 2018;80(3):864–873. doi: 10.1002/mrm.27077 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Grist JT, McLean MA, Riemer F, et al. Quantifying normal human brain metabolism using hyperpolarized [1–13C]pyruvate and magnetic resonance imaging. NeuroImage. 2019;189:171–179. doi: 10.1016/j.neuroimage.2019.01.027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Lee CY, Soliman H, Geraghty BJ, et al. Lactate topography of the human brain using hyperpolarized 13C-MRI. NeuroImage. 2020;204:116202. doi: 10.1016/j.neuroimage.2019.116202 [DOI] [PubMed] [Google Scholar]
  • 18.Gordon JW, Autry AW, Tang S, et al. A variable resolution approach for improved acquisition of hyperpolarized 13 C metabolic MRI. Magn Reson Med. 2020;84(6):2943–2952. doi: 10.1002/mrm.28421 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Chen H, Autry AW, Brender JR, et al. Tensor image enhancement and optimal multichannel receiver combination analyses for human hyperpolarized 13 C MRSI. Magn Reson Med. 2020;84(6):3351–3365. doi: 10.1002/mrm.28328 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Kim Y, Chen HY, Autry AW, et al. Denoising of hyperpolarized 13 C MR images of the human brain using patch-based higher-order singular value decomposition. Magn Reson Med. Published online June 25, 2021. doi: 10.1002/mrm.28887 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Ma J, Park JM. Super-Resolution Hyperpolarized 13C Imaging of Human Brain Using Patch-Based Algorithm. Tomography. 2020;6(4):343–355. doi: 10.18383/j.tom.2020.00037 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Gordon JW, Chen HY, Autry A, et al. Translation of Carbon-13 EPI for hyperpolarized MR molecular imaging of prostate and brain cancer patients. Magn Reson Med. 2019;81(4):2702–2709. doi: 10.1002/mrm.27549 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Pruessmann KP, Weiger M, Börnert P, Boesiger P. Advances in sensitivity encoding with arbitrary k-space trajectories. Magn Reson Med. 2001;46(4):638–651. doi: 10.1002/mrm.1241 [DOI] [PubMed] [Google Scholar]
  • 24.Zhu Z, Zhu X, Ohliger MA, et al. Coil combination methods for multi-channel hyperpolarized 13C imaging data from human studies. Journal of Magnetic Resonance. 2019;301:73–79. doi: 10.1016/j.jmr.2019.01.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Zhang Y, Brady M, Smith S. Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging. 2001;20(1):45–57. doi: 10.1109/42.906424 [DOI] [PubMed] [Google Scholar]
  • 26.Mammoli D, Gordon J, Autry A, et al. Kinetic Modeling of Hyperpolarized Carbon-13 Pyruvate Metabolism in the Human Brain. IEEE Trans Med Imaging. 2020;39(2):320–327. doi: 10.1109/TMI.2019.2926437 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Park I, Lupo JM, Nelson SJ. Correlation of Tumor Perfusion Between Carbon-13 Imaging with Hyperpolarized Pyruvate and Dynamic Susceptibility Contrast MRI in Pre-Clinical Model of Glioblastoma. Mol Imaging Biol. 2019;21(4):626–632. doi: 10.1007/s11307-018-1275-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Jezzard P, Balaban RS. Correction for geometric distortion in echo planar images from B0 field variations. Magnetic Resonance in Medicine. 1995;34(1):65–73. doi: 10.1002/mrm.1910340111 [DOI] [PubMed] [Google Scholar]
  • 29.Miller JJ, Lau AZ, Tyler DJ. Susceptibility-induced distortion correction in hyperpolarized echo planar imaging. Magnetic Resonance in Medicine. 2018;79(4):2135–2141. doi: 10.1002/mrm.26839 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Chen HY, Gordon JW, Bok RA, et al. Pulse sequence considerations for quantification of pyruvate-to-lactate conversion k PL in hyperpolarized 13 C imaging. NMR in Biomedicine. 2019;32(3):e4052. doi: 10.1002/nbm.4052 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Larson PEZ, Chen H, Gordon JW, et al. Investigation of analysis methods for hyperpolarized 13C-pyruvate metabolic MRI in prostate cancer patients. NMR in Biomedicine. 2018;31(11). doi: 10.1002/nbm.3997 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supinfo

Supporting Information Table S1. Mean and standard deviation (SD) of pyruvate-to-lactate conversion (kPL) and lactate-to-pyruvate area-under-curve (AUC) ratios for multi-resolution and constant-resolution C13 pyruvate data in the brain for six subjects. Multi-resolution kPL and AUC ratio means and SDs were significantly higher (p < 0.05) than constant-resolution kPL and AUC ratio means and SDs. Average differences for multi-resolution as compared to constant-resolution were: 19% and 70% increase in kPL mean and SD; 22% and 93% increase in AUC ratio mean and SD.

Supporting Information Table S2. Pyruvate-to-lactate conversion (kPL) values for multi-resolution and constant-resolution C13 pyruvate data from four regions in the brain. Single voxels near arteries, veins, in white matter, and in gray matter were selected manually. Multi-resolution kPL was significantly higher for voxels near arteries and veins and voxels in white matter (p < 0.05), showing a decrease in partial volume spillover of vascular pyruvate. There were inconsistent differences for gray matter voxels. Subject 4 values were an average of two C13 pyruvate injections in the same exam and stars indicate statistical significance (p < 0.05).

Supporting Information Table S3. Lactate-to-pyruvate area-under-curve (AUC) ratios for multi-resolution and constant-resolution C13 pyruvate data from four regions in the brain. Single voxels near arteries, veins, in white matter, and in gray matter were selected manually. Multi-resolution AUC ratios were significantly higher for voxels near arteries and veins and voxels in white matter (p < 0.05), showing a decrease in partial volume spillover of vascular pyruvate. There were inconsistent differences for gray matter voxels. Subject 4 values were an average of two C13 pyruvate injections in the same exam and stars indicate statistical significance (p < 0.05).

NIHMS1810382-supplement-supinfo.docx (24KB, docx)

RESOURCES