Fig. 1) (a) Global lactate-to-pyruvate ratio as a function of age with associated Spearman correlation and p-value (Spearman = -0.685, p = 0.001). (b) Global bicarbonate-to-pyruvate ratio as a function of age with associated Spearman correlation and p-value (Spearman = -0.586, p = 0.008). Error analysis is ongoing.

Fig. 2) Regional regressions of the lactate-to-pyruvate ratio as a function of age were used to predict an approximate rate of change per decade relative to the mean age of 46. These rates of change were then linearly mapped to a 3D rendering of the human brain.

(A) Representative ¹³C spectra showing HP [1-¹³C]pyruvate and HP [1-¹³C]lactate production from a region containing the CA1 area of the hippocampus (red square). (B) HP [1-¹³C]lactate/pyruvate ratios were significantly lower in female GLUT3 cKO compared to GLUT3 WT (p=0.0282). (C) HP [1-¹³C]lactate/pyruvate ratios were not significantly different between male GLUT3WT and GLUT3cKO.

(A) 3D rendering of in vivo T₂w MRI showing the entire brain (grey), hippocampus (red), thalamus (yellow) and ventricle (blue). (B) In vivo T₂w images revealed significantly smaller brain, hippocampus and thalamus in female GLUT3 cKO compared to WT. (C) No changes were observed in males. (D) Ex vivo T₂w images showing the hippocampus and CA1 region (arrow). CA1, hippocampus and total brain volumes calculated from ex vivo T₂w images were significantly smaller in female GLUT3 cKO compared to WT.

Figure 4: Area-under-the-curve ratios of hyperpolarized lactate over pyruvate for spleens (a) and kidney regions (b). Tumor metabolic volumes are calculated for spleens (c) and spleen sizes are measured (d). Significant differences indicated with dotted line p<0.01 or striped line p<0.05. The different resolution reached in ¹³C imaging is indicated.

Figure 2: a. ROI are selected in ¹H-RARE images, spleen indicated. b) ¹³C images of hyperpolarized lactate (orange) and pyruvate (blue) show high levels of lactate in kidneys. Calculation of integrals over time of these metabolites results in the curves in c.) Analysis of AUC ratios gives an unique insight in exchange rate between of HP pyruvate and HP lactate.

Effect of exercise in HP 13C MRS. Time-averaged 13C spectra over the first 90 seconds after bolus injection of HP [1-13C]pyruvate from a representative participant (a) at rest, (b) during exercise, and (c) recovery. (d) Spatially-averaged perfusion measured by 1H ASL from all the participants. (e) Total 13C signals (tC). Area under the curves (AUCs) for (f) [1-13C]pyruvate, (g) [1-13C]lactate, (h) [1-13C]alanine, and (i) [13C]bicarbonate, normalized by the tC AUC. AUCs of (j) alanine and (k) bicarbonate, normalized by the lactate AUC. (*: p<0.05, **: p<0.01).

Experimental setup and study protocol. (a) Positioning calf muscle in a 13C/1H dual-frequency RF coil prior to connect anterior part of the coil. (b) Calf muscle was wrapped by the flexible posterior 13C receive arrays (channel #1-4). (c) Localized single-shot fast spin echo 1H images acquired using the RF coil. Blue region indicates the prescribed axial slab for 13C MRS (10-cm thick). (d) Dynamic 13C MRS was acquired at three metabolic states: rest, exercise, and recovery.

Hyperpolarized δ-[1-¹³C]gluconolactone informs on early response to TMZ treatment in vivo. Top: representative metabolic maps from a BT88-bearing rat pre-treatment with TMZ. T2 weighted MRI (A), heatmap of SNR of hyperpolarized δ-[1-¹³C]gluconolactone (B) and heatmap of 6PG/δ-[1-¹³C]gluconolactone ratio (C). Bottom: metabolic maps from a BT88-bearing rat post-treatment with TMZ. T2 weighted MRI (D), heatmap of SNR of hyperpolarized δ-[1-¹³C]gluconolactone (E) and heatmap of 6PG/δ-[1-¹³C]gluconolactone ratio (F). Tumor is delineated in white.

Hyperpolarized δ-[1-¹³C]gluconolactone can monitor response to temozolomide (TMZ) in glioma cells. Representative ¹³C-MRS spectral array (A) and summed ¹³C spectra (B) showing 6PG production from hyperpolarized δ-[1-¹³C]gluconolactone in live U87 cells. Effect of TMZ on 6PG production in U87 (C) and GS2 (D) cells (N=3).

Figure 2: (A) shows small to moderate sized ventilation defects in a symptomatic patient with consistently normal spirometry. (B) shows relatively homogeneous ventilation in a highly symptomatic patient.

Table 3: Summary of subjects’ MRI classification, spirometry and ACQ7 scores. MRI classified based on radiologists’ reports (normal = no or minor ventilation defects, abnormal = substantial ventilation defects). Note: ACQ7 score was available for 24 patients within 3 months of the scan.

Figure 5. ¹²⁹Xe MRI/MRS and RBC oscillation imaging for 2 subjects with CpcPH and no extensive obstructive or interstitial disease. The conventional algorithm classifies Subject A as no PH, and subject B as precapillary PH. Oscillation imaging in both subjects indicates regions of both low and high oscillations. Thus, imaging appears to correctly indicate the presence of both pre- and postcapillary PH.

Figure 4. ¹²⁹Xe gas exchange and RBC oscillation imaging for test subjects classified by RHC as pre-capillary PH. Subjects A and B exhibit RBC amplitude oscillations of 5.6%, and 5.4%, and are both properly classified as PH-pre. These subjects also exhibit regions of low RBC oscillations on imaging. Subject C and D have RBC oscillation amplitudes of 10.9% and 9.3% leading the basic algorithm to incorrectly classify both classified as no PH. These subjects have extensive parenchymal disease and oscillation imaging is heterogeneous, suggesting PH may be in proportion to lung disease.

Figure 1. Pulse sequence diagram. The phase difference of the two 90° pulses was chosen to be 0 at the given off-resonance frequency Δf. Fermi-shaped pulses are used for selectively creating spatially modulated magnetization in the ¹²⁹Xe dissolved phase while providing a sufficient amplitude integral to achieve 90° flip angles within the dissolved phase T₂*. The whole sequence is repeated as a control acquisition with inverted off-resonance frequency and STEAM de- and rephasing gradients rotated by 90°.

Figure 2. Exemplary line-broadened spectra after off-resonance correction and phasing for the FID signal from a) a healthy volunteer and b) a COPD patient after excitation at +3485 Hz. Real part of FID signal shown in orange, STE signal in blue and rescaled by factor 100. In a) as expected for pure TP excitation, an approximate phase difference of π is found. In b) at later times a noticeable phase change is observed, possibly due to signal contamination or motion. Control acquisitions are shown for the HV in c) and the COPD patient in d). Only small signals are observed in the STE pathway.

Figure 1: Qualitative results of ventilation comparison between control group and long-haul COVID-19 patients

Control group participants S01 and S02 have a VDP of 2% and 1%, respectively, while COVID-19 patients S15 and S16 have a VDP of 3% and 2%, respectively.

Figure 2: MRI ventilation heterogeneity measurements

Patients with long-haul COVID-19 have increased VDP (A), short run emphasis (B), high gray-level run emphasis (C) and short run, high gray-level emphasis (D) extracted from ¹²⁹Xe MRI, compared to control group participants.

Figure 2. Representative max intensity projections in the coronal plane from one mouse. a) 3D radial, 25% percent sampled. b-d) 3D spiral (FLORET), with each spiral arm durations equal to b) 1 x T₂* readout, c) 2 x T₂* readout, and d) 3 x T₂* readout.

Figure 4. Box plots summarize the values of a) estimated initial polarization, b) Signal-to-noise ratio, and c) the corrected signal-to-noise ratio for each of the four scans. Descriptions for the calculation of SNR and the corrected SNR are detailed in the methods section. * ≡ significant difference (p = 0.0008).

Fig. 3) The visual effect of the various models trained on the k-space removal size 128x128 dataset on a healthy subject (rows 1-2) and one with visible ventilation defects (rows 3-4). Regions of interests are highlighted to show differences in image texture, image noise, and feature sharpness. Row 2 indicates that edges are sharpened while decreasing background noise. Row 4 indicates the recovery of the ventilation defect.

Fig. 4) Benchmark results (PSNR/SSIM/SNR) of all experiments. For models trained with datasets generated from k-space under-sampling (columns 2,3), improvements are seen in all three metrics compared to the bicubic up-sampling method on the low-resolution image. Moreover, SNR of all model outputs are higher than that of the original ground truth. However, for models trained with datasets generated from bicubic down-sampling of the high-resolution image (column 1), SNR is decreased because the models are able to accurately reproduce the noise pattern.

Fig. 1. (a) Uncertainty in ADC measurements. (b) A clear decrease in the relative uncertainty, 𝝐_ADC, is observed with increasing SNR₀. (c) Shows the decrease of 𝝐_ADC as N_ph decreases across the ADC range, where this decrease is more pronounced for very small and very large ADC values. (d) Shows the increase of 𝝐_ADC as N_b increases across the ADC range with global minima at N_b≤4. However, for large ¹²⁹Xe ADC (>0.06 cm²/s) using only 2 b-values increases the relative error significantly. (e) Shows the global minima of at M_z(R)=20% across the ADC range.

Fig. 4. In vivo HP ¹²⁹Xe ADC mapping. (a) b₀ image alongside with (b) ADC maps of a 6, 18 and 30 years old healthy subjects [Note: The b₀ images and the ADC maps were interpolated for visualization only]. (c) Corresponding ADC histograms obtained from the subjects imaged in a. Mean ADC increases noticeably with aging in these subjects.

Figure 4. Lobar weighted entropy analysis for the asthmatic subjects shown in Figure 3. A) the global weighted entropy decreased from 113.1 to 45.8 after BD, while the left lung and the right lower lobe contributed the most to the weighted entropy decrease. B) the global weighted entropy decreased from 53.5 to 48.7 after BD, while all the lobes except the left lower lobe contributed to the weighted entropy decrease. C) the global weighted entropy increased from 26.8 to 33.7 after BD, while the left lower lobe and right lower lobe contributed to this weighted entropy increase.

Figure 3. A) An asthma patient with FEV1_%pred=39.2, FVC_%pred = 88.5, and VDP = 25.4% at baseline. The weighted entropy was 113.1. B) An asthma patient with FEV1_%pred=86, FVC_%pred = 96, and VDP = 11.4% at baseline. The weighted entropy was 53.5. C) An aasthma patient with FEV1_%pred=95.6, FVC_%pred = 102.9, and VDP = 6% at baseline. The weighted entropy was 26.8.

Figure 3. Example coronal slices of the UNet and VNet combined ³He and ¹²⁹Xe trained segmentations for three cases with different diseases compared to the expert segmentations. DSC and Avg HD values are given for each case.

Figure 4. Correlation and agreement analysis of lung volumes for 75 testing set cases compared to expert segmentations for combined ³He and ¹²⁹Xe a) VNet and b) UNet models.

Table 1. Participant baseline demographics, pulmonary function and imaging information.

Figure 2. Logistic regression analysis of individual variables in A) clinical and B) MRI model. The above graphs represent the predictive power of individual variables used in each of the models. Clinical model significantly improved from adding lung volume variables. Individual texture features outperformed standard variables calculated from MRI, namely ADC and VDP.

Fig. 4: Volunteer 1: Gas, tissue, blood and its corresponding ratio maps. The heart is visible in the MRSI, but not radial blood maps. Ratio maps in percent are scaled to 2.5%, 0.9% and 70% for tissue/gas, blood/gas and blood/tissue ratios, respectively.

Fig. 3: Representative spectrum from volunteer 1 (x=13, y=16, z=7). Besides spectral zero-filling to 256 points and spatial zero-filling by a factor of 2, no further post-processing steps (such as apodisation or denoising) were performed. Some Gibb’s ringing is visible around the gas peak due to the longer T₂^*. This exemplifies the good spectral quality and excellent SNR.

Figure 1. The ¹²⁹Xe/¹H switched frequency high pass birdcage coil with the outer cylinder removed.

Figure 3. ¹²⁹Xe spoiled gradient echo (SPGR) coronal images of a hyperpolarized ¹²⁹Xe gas filled Tedlar® bag.

Figure 4: The RBC signal (red line) is divided by the fit in Fig. 2 (b) and subtracted by the mean signal value to obtain the fractional signal change for RBC over time. Overlaid in black is the ECG signal which was recorded during the MR acquisition. It is clear from the ECG overlay that the RBC signal is at maximum and minimum during ventricular systole and diastole, respectively.

Figure 2: Shown in (a) are the RBC and TP signal amplitudes as a function of breath-hold time. The steady-state RBC signal amplitude is shown in (b) along with a quadratic fit for the purpose of signal normalisation.

Figure 2: Depolarization maps in a healthy rat and corresponding boxplots for 500 ms TP saturations with increasing flip angles.

Figure 3: Depolarization maps for RBC saturations in a healthy rat, comparing equivalent flip angle per ms saturations with multiple, shorter pulses (< 0.3 ms delay in-between). For example, the first column is: one 100 ms pulse (flip angle 18,000°) vs. two 50 ms pulses (flip angle 9,000° each) vs. four 25 ms pulses (flip angle 4,500° each).

Figure 4: All test-retest correlation plots for functional maps (e.g, TV, lower, blue colormap and , red colormap) in two back-to-back repeats in a healthy nonsmoker and a severe COPD.

Figure 3: Images of HP 129Xe xenon dissolved in the grey matter for both volunteers at the acquired slice thickness of 2 cm.

Figure 2: Average spectrum of all the spatially resolved 1101 RF-acquisitions from Volunteer 1 (Male 37 years).

Figure 1: Example single slice ¹²⁹Xe images from Healthy Volunteer 5. (Top) Magnitude (b=0), ADC map, and Lm_D map for data acquired at 1.5T. (Bottom) The equivalent images acquired at 3T. The white arrows denote regions around the diaphragm where magnetic susceptibility differences can be observed that are not present at 1.5T.

Figure 2: (a) Bland-Altman comparison of mean global ¹²⁹Xe ADC (b=12) at 3T and 1.5T for the five healthy volunteers. A mean bias of 6.3% (solid line) towards 3T was observed with a 95% confidence interval of -2.7% to 15.3% (dotted lines). (b) Bland-Altman comparison of mean global ¹²⁹Xe Lm_D with a mean bias of 2.2% towards 3T and a 95% confidence interval of -4.0% to 8.5%. (c) Bland-Altman comparison of mean global ¹²⁹Xe alpha heterogeneity index with a mean bias of -0.2% towards 3T and a 95% confidence interval of -5.3% to 5.0%.

Figure 1. CT renderings of the spine (left column) and the lungs (right column) in a rib-tether rabbit model immediately after surgery (6 weeks of age, top row) and at 22 weeks post surgery (28 weeks of age, bottom row) of age. The successful model implementation resulted in a spinal deformation that highly restricted the expansion of the right lung during maturation.

Figure 2. GP signal dynamics during multi-breath 1D projection acquisitions aggregated for the left and right lung. (a) In the age-matched control rabbit, both lungs are ventilated symmetrically throughout the respiratory cycle. (b) In the rib-tethered rabbit, the restricted right lung is more poorly ventilated than the left lung, most likely due to decreased compliance.

Fig. 1: (A) CAD drawings of the SEOP cells for different Rb source-sink distributions (highlighted): (i) the SEOP cell with a Rb pool located near the inlet; (ii) the SEOP cell with Rb on all internal surfaces of the cell, and (iii) the SEOP cell with Rb located in the presaturator. (B) Corresponding normalised Rb density distributions, [Rb]/[Rb]_sat, for laser power =160W and gas flow rate =2000sccm. Rb density is normalised to the saturation density given by the Killian formula, [Rb]_sat=2.70x10¹³cm^-3.

Fig. 5: (a) Normalised Rb density, [Rb]/[Rb]_sat, versus position along the presaturator for a gas flow rate of 900sccm (blue), 1100sccm (orange), 1400sccm (yellow), 1700sccm (purple) and 2000sccm (green). (b) Length of presaturator required for [Rb]=0.99*[Rb]_sat versus flow rate. The presaturator is cylindrical, with a diameter of 6mm and length of 73.1cm. The Rb source covers 25% of the total internal surface area of the presaturator.

Figure 3. Comparison of corrected images using template-based approach and its respective direct correction method (RF-depletion mapping and N4ITK). These examples illustrate 1) direct application of N4ITK tends to be similar or more aggressive than template-based bias field correction and 2) Template based bias field correction based off of RF-depletion mapping achieves similar levels of correction compared to direct application of RF-depletion mapping derived from the individual subject.

Figure 2. Illustration of template-based correction. First the template mask is registered to the subject mask, and the same transformation is applied then to the bias field template. Then, inpainting is used to fill in dark areas at the boundary caused by the transformed volume moving partially out of the FOV.

Maps representing ratio of depolarized/control image in a wild-type mouse, with increasing delay times (1.3, 2.3, 3.4, 7.3, 10.633, 17.3, 37, 64) and decreasing pulses (50, 40, 30, 20, 15, 10, 5, 3), for a same total of saturation time. Highest depolarization reached is about 30%, at a 3.4 ms delay.

Top: plots representing signal loss per pulse in the gas peak at a range of delay times, in a wild type mouse (saturated at dissolved), and in a transgenic mouse (saturated at tissue and RBC).

Bottom plots representing mean depolarization at each delay time wild type and transgenic models.

Five representative examples of CNN correction performance. Images show cyan ³He images overlaying ¹H images. Top row: images with random affine transformations applied. Middle row: proposed registration generated by multi-input UNet. Bottom row: original co-registered images obtained through semi-automated, landmark based registration.

Participant demographics, pulmonary function tests and imaging measurements. Values reported as mean (standard deviation). BMI = body mass index, PFT = pulmonary function tests, FEV₁ = forced expiratory volume in 1 second, FVC = functional vital capacity, RV = residual volume, TLV = total lung volume, VDP = ventilation defect percent

Figure 2. (A) HP ¹²⁹Xe brain images of representative healthy volunteer acquired using a GRE imaging 9 s into the breath-hold without an initial depolarization pulse. (B) Standard SNR deviation map calculated based on images (A). (C) HP GRE ¹²⁹Xe brain images acquired 9 s after the initial depolarization pulse. (D) Standard SNR deviation map calculated for (C) images. It can be clearly seen that the application of the initial depolarization pulse yields lower signal variability

Figure 4. Comparative box-chart of the image SNR. The black box represents the SNR of the GRE images acquired 9s into the breath-hold without an initial depolarization radiofrequency pulse. The red box represents the SNR of the GRE images acquired 5s after the initial depolarization pulse. The blue box corresponds to the SNR of the GRE images acquired 9s after the initial depolarization pulse.

Figure 1: Representative b₀ images for (a) 2D-GRE, (b) 2D-spiral and (c) 3D-FLORET sequences. Spiral artifacts present near the edges of high signal regions are indicated with arrows. Scan time was 7 s (0.21 ms per voxel) for 2D-spiral and 16 s (0.11 ms per voxel) for 3D-FLORET compared to 15 s (0.73 ms per voxel) for 2D-GRE.

Figure 3: ADC maps for (a) 2D-GRE, (b) 2D-spiral and (c) 3D-FLORET, shown no meaningful difference in mean ADC.

Figure 1. ¹²⁹Xe MRI fractal analysis outline.

¹²⁹Xe MRI static ventilation was segmented into 5 clusters of signal intensity ranging from signal void or ventilation defects to hyperintense signal (Cluster5) (A) and the box-counting method was used to calculate fractal dimension (B). Fractal dimension was calculated for each MRI signal intensity cluster (C).

Figure 2. ¹²⁹Xe MRI fractal dimension measurements.

(A) Fractal dimension (D), R² and root-mean-squared error (RMSE) for all MRI signal clusters. (B) ¹²⁹Xe MRI ventilation (cyan) co-registered to anatomical ¹H (grey-scale) for representative participants with asthma with associated fractal dimensions. (C) Log-log plots showing fractal dimensions for defect regions (Cluster1), Cluster2 and Cluster4 for representative participants.

Figure 1. Acquisition scheme for each metabolite and the proposed imaging sequence. (A) For each metabolite, two acquisitions were conducted during each cardiac cycle, which are at end systole and end diastole, respectively. (B) Multi-echo images for hyperpolarized [¹³C]bicarbonate, [¹³C]lactate, [1-¹³C]alanine and [1-¹³C]pyruvate are acquired in order with the interval of 1 R-R for each timepoint. Images from in total 16 timepoints are acquired.

Figure 3. Dual-phase hyperpolarized [1-¹³C]pyruvate cardiac imaging in SA plane. (A) ¹H cardiac images in SA plane with the trigger delays of 385 ms (end systole) and 791 ms (end diastole) respectively from a healthy subject. (B) Dynamic change of hyperpolarized [¹³C]bicarbonate, [1-¹³C]lactate, [1-¹³C]alanine and [1-¹³C]pyruvate in two phases from 15 s to 35 s post the injection of pyruvate.

In vivo pyruvate-to-lactate conversion is significantly altered throughout tumor development, regression, and recurrence. Individual nLac values, measured with hyperpolarized ¹³C MRS, are plotted as a function of time for control mice (red circles) and untreated tumor-bearing mice (blue squares) during tumor development (A). Individual nLac values are plotted for treated tumor-bearing mice (green triangles) during tumor regression and recurrence (B). Individual nLac values are plotted across the entirety of tumor progression (C). *p < 0.05, **p < 0.01, ***p < 0.001.

Percent change of nLac, but not tumor volume, is significantly altered during tumor regression and recurrence. Repeated measures of tumor volume are acquired over time in treated mice with anatomic MRI (A) and hyperpolarized MRS (B). At each time-point, volume and nLac were normalized to their initial value following treatment and plotted as percent change over time (C). *p < 0.05, ***p < 0.001.

Figure 4. In vivo CSI of hyperpolarized [1-¹³C]fumarate metabolism in an acetaminophen-induced hepatitis mouse. (a) Representative ¹³C NMR spectrum of hyperpolarized [1-¹³C]fumarate and its metabolite at the liver. (b) map of hyperpolarized [1-¹³C]fumarate CSI signal intensity overlaid on an anatomical ¹H MRI image. (c) map of hyperpolarized [1-¹³C] & [4-¹³C]malate. (d) parametric map of the malate/fumarate ratio; a biomarker of cellular necrosis.

Figure 2. (a) Preparation of hyperpolarized [1-¹³C]fumarate by trans-alkenylation with parahydrogen. (b) ¹³C NMR of hyperpolarized [1-¹³C] fumarate using different ¹H-to-¹³C spin order transfer pulse sequences.

Figure 1. Lactate and alanine flux within N1S1 and MCA-RH7777 tumors as measured with h¹³C MRI. Lactate/pyruvate (A) and alanine/ pyruvate (B) ratios measured in N1S1 tumors before and 72 hours after RFA versus sham (control) treatments. Lactate/ pyruvate (C) and alanine/ pyruvate (D) ratios in MCA-RH7777 tumors before and 72 hours after RFA versus sham (control) treatments. *, p<0.05, ***, p<0.001.

Figure 2. N1S1 and MCA-RH7777 tumor glycolysis-related mRNA expression following treatment. (A) Glycolysis-related gene expression pattern in N1S1 and MCA-RH7777 cell line. (B) PFKFB3 expression in N1S1 tumor (T), adjacent normal liver (N) and RFA/sham site (R), for RFA and sham (control) arms. Quantification of band intensity is presented as % of relative densitometry normalized to the CYPA gene. *, p<0.05.

nLac maps overlaid on top of T2w images at baseline (left) or 8 days into systemic treatment (right). All high nLac voxels were confined to the tumor. Measured nLac values were reduced 31% 8 days after the onset of systemic therapy.

Baseline area under the curve images for hyperpolarized pyruvate (left) and lactate (right) overlaid on top of T2w images of a large left-sided ATC tumor identified with the white arrows on slice 2. Pyruvate signal is primarily observed in the vascular anatomy while lactate signal is localized to the tumor.

Flowcharts of (a) the auto-focusing correction method and (b) time-segmented NUFFT correction with self-estimated field map method.

Reconstruction results of brain data with pyruvate signal. The images are processed (a) without off-resonance correction, (b) auto-focusing correction, and (c) iterative NUFFT with self-estimated field map correction. We compared dynamic images at five time points and area-under-the-curve (AUC) images. The green arrows show the comparison between different methods and illustrate method(c) has better correction performance. A line profile is chosen for intensity comparison.

Conversion of [5-13C,4-2H2,5-15N­]-L-Glutamine to [5-13C,4-2H2]-L-Glutamate with glutaminase in tumor. Red, orange, blue represent sites of 13C, 2H and 15N enrichment, respectively.

A T2-weighted 1H MRI of mouse, with tumor (right) highlighted as regions of sagittal slab B Dynamic of conversion of glutamine to glutamate in the tumor with vehicle. C Dynamic of conversion of glutamine to glutamate in the tumor with glutaminase inhibitor CB-839

Figure 1. Imaging comparison between PDXs. Images show differences in tumor morphology (A1-A4), tumor cellularity (B-1-B4) and tumor glycolysis in tumors (C-1-C-4). D) Bar graph shows the mean %cystic component and F) %necrosis component. G) Bar graph of the ADC mean calculated for all tumor slices. H) Bar graph of the lactate to pyruvate conversion rate in tumors. a.u.: arbitrary units.a.u.: arbitrary units. The significance of the Dunn’s post-test. Red dots in bar graphs indicate passage 3 and 4 of PDX047.

Figure 2. Imaging comparison between PDX054 and its xenograft XEN054. Images show differences in tumor morphology (A1-A4), tumor cellularity (B-1-B4) and tumor glycolysis in tumors (C-1-C-2). D) Bar graph shows the mean %cystic component and F) %necrosis component. G) Bar graph of the ADC mean calculated for all tumor slices. H) Bar graph of the lactate to pyruvate conversion ratio in tumors. a.u.: arbitrary units. The significance of the Mann-Whitney test is indicated above the plots. All p < 0.05 indicated statistical significance.

Figure 1. T₂* measurement of HP ¹³C-labeled metabolites for human heart. (A) Short-axis ¹H MRI of heart from a healthy subject. (B) Dynamic changes of [¹³C]bicarbonate, [1-¹³C]lactate and [1-¹³C]pyruvate from 25 s to 55 s after the injection of HP pyruvate. (C) The first 6 echoes acquired at 25 s post the injection for [¹³C]bicarbonate, [1-¹³C]lactate and [1-¹³C]pyruvate. (D) Changes of ¹³C signals within LV, RV and Myo along the echo time in log scale at 25 s post the injection.

Figure 2. T₂* measurement of HP ¹³C-labeled metabolites for human brain. (A) Axial ¹H MRI of brain from a healthy subject, according to which GM, WM and CSF were segmented. (B) Dynamic changes of [¹³C]bicarbonate, [1-¹³C]lactate and [1-¹³C]pyruvate from 11 s to 36 s after the injection of HP pyruvate. (C) The first 6 echoes acquired at 26 s post the injection for [¹³C]bicarbonate, [1-¹³C]lactate and [1-¹³C]pyruvate. (D) Changes of ¹³C signals within GM, RM and CSF along the echo time in log scale at 26 s post the injection.

Figure 2. Kinetic rate maps for pyruvate to lactate conversion, using 15 × 15 mm² lactate signals with 7.5 × 7.5 mm² pyruvate and synthetic 15 × 15 mm² pyruvate images thresholded to SNR >5 and <30% fit error, along with reference proton images. The constant-resolution k_PL map shows smoothing of the kinetic rates in the brain as compared to the multi-resolution k_PL map.

Figure 1. Hyperpolarized ¹³C pyruvate, lactate and bicarbonate signals summed over 60 seconds with reference proton images. The synthetic 15 × 15 mm² pyruvate data set was obtained by convolving the 7.5 × 7.5 mm² pyruvate data with an averaging kernel. The red arrows point out the high pyruvate signal in the transverse sinus for the third slice and in the superior sagittal sinus for the sixth and seventh slices.

(A) Discretized shaped pulses on ¹H and ¹⁵N channels. The color-coding and the height of the bars represent the phase (Φ_i) and the amplitude of the pulse points respectively. The length of the pulse points (τ_i) is also an optimization variable. (B) Pulse amplitude on proton channel for the ¹H to ¹⁵N full transfer. (C) Pulse amplitude on the ¹⁵N channel. Despite the rapidly varying waveforms the predicted behavior was well preserved.

Time course for ¹⁵N to ¹H partial polarization transfer in a hyperpolarized [¹⁵N₂]urea phantom. The broad urea peak is the result of the degraded magnetic field homogeneity caused by injection of the hyperpolarized urea. The residual water peak is due to insufficient water suppression

Figure 2: Representative dynamic cerebral ¹³C MRS acquired after a bolus infusion of [1-¹³C]lactate (lb=10Hz). The summed signal from the first 90s post-infusion is plotted in red. The vertical scale was normalized to the height of the summed HP lactate peak. In both healthy and stroke animals, the HP lactate was converted into [1-¹³C]pyruvate (171.1 ppm), [1-¹³C]alanine (176.7 ppm) and [¹³C]bicarbonate (161.2 ppm). The signal observed at 177.7 ppm (*) is an impurity from the stock lactate solution.

Figure 3: Metabolite ratios following HP lactate infusion at 1h post-surgery or post-reperfusion. Data are represented as mean ± SD. (a) Following infusion of [1-¹³C]lactate, the normalized pyruvate-to-lactate ratio (cPLR) was significantly lower in the MCAO group compared to both sham and MCAO+TEMPOL groups. (b) A trend towards lower alanine labeling (cALR) was observed after stroke, with a significant difference between MCAO+TEMPOL and sham. (c) The bicarbonate-to-lactate ratio (cBPR) tends to decrease after stroke.

(A) ¹³C polarization losses as a function of the sample vertical position inside the polarizer while using a traditional DNP probe (blue and black circles) and our new probe containing permanent magnets (red circles).The gray shaded area represents the area covered by permanent magnets in the new probe. (B) Measured (black line) and calculated (red line) polarizer magnetic field profile. (C) Calculated total (green line) magnetic field: superconductive magnet (red dotted line) + permanent magnets (blue dotted line).

Technical drawings of the CFP (A) and zoomed section of the T-valve indicating inner (orange arrow) and outer (cyan arrows) flow directions (B), dynamic sealing (C) and sample threaded vial (D). Numbers indicates the most important components of the device: quick release connection (1), T-valve (2), one-way valve (3), dynamic sealing (4), vial top part (5), vial bottom part (6), outer-lumen to inner-lumen transition (7), black PEEK outer-lumen (8), red PEEK inner-lumen (9), dynamic sealing silicone o-ring (10), laser welded joint (11), vial PTFE o-ring (12), nozzle (13).

Figure 2. Lactate to pyruvate (L/P) ratio maps at baseline, and week 9 and 18 in rats fed the MCD diet. b) Graph of L/P ratio values at baseline, and week 9 and 18 in rats fed the MCD diet. T-tests were performed to compare changes between time points.

Figure 1. Fat signal fraction (FSF) maps at baseline, and week 9 and 18 in rats fed the MCD diet. b) Graph of FSF values and c) body weight values at baseline, and week 9 and 18 in rats fed the MCD diet. T-tests were performed to compare changes between time points.

Data acquisition and constrained reconstruction scheme of CoUNFOLD. P stands for a pixel in the full FOV image, A stands for an aliased pixel in the undersampled (R stands for the ratio) image with reduced FOV, A’ is the estimated aliased pixel that is calculated under the same undersampling scheme as A from the CoUNFOLD reconstructed image. The solid and dashed lines represent, respectively, the acquired and skipped phase steps in k-space.

The time series at the same voxel location from different undersampling rates. (a) The sampled metabolite series ('x') at SNR=30 are fluctuating along the noise free series (dash lines), while the estimated series (solid) from pharmacokinetic model matches the noise free series with slight offset. (b) The undersampled pyruvate series (blue 'x') at SNR=30 fluctuate drastically around the noise-free series(dashed lines) and the lactate series(red 'x') are severely deviated, but the CoUNFOLD restored series (solid) match the noise free series only with moderate offset.

Figure 1. Examples of time-resolved 13C dynamic spectra from a 25 mm slice of rat brain from a CA model (B) and a sham control (C). Hyperpolarized 13C MR spectroscopy data revealed that the cardiac arrest model produced a high level of pyruvate-to-lactate conversion while the sham control had a smaller level of pyruvate-to-lactate conversion compared to the CA model.

Figure 2. The ratio of lactate to pyruvate in the CA model was statistically higher than that in the sham control (P=0.025).

Fig. 4: Reconstructed metabolic maps for the in vivo 2D rat brain imaging data at selected time points with (a) fully sampled data (b) Two-fold randomly undersampled data using direct NUFFT (c) Two-fold randomly undersampled data using L+S reconstruction with one L+S process.

Fig. 2: Reconstructed metabolic maps for the digitally simulated 2D rat imaging data at selected time points with (a) fully sampled data (b) Two-fold randomly undersampled data using direct NUFFT (c) Two-fold randomly undersampled data using L+S reconstruction with one L+S process.

Figure 1. Representative spectra of renal [1-13C]pyruvate metabolism. PEP-CK inhibition by 3-MPA results in markedly lower aspartate and malate signals. Abbreviations: Mal – malate, Asp – aspartate, Fum – fumarate.

Figure 2. Effects of fasting and 3-MPA treatment on renal metabolism of hyperpolarized 1-13C pyruvate to bicarbonate, aspartate, lactate and alanine. The factors associated with the noted 2-way ANOVA p values are indicated in parentheses.

Fig. 2: (A) Representative EM images of mitochondrial structure in non cancer (GM) ATM +/+ & ATM-/- & DLBCL cell lines, HLY-NT & sh-HLY. ATM unmutated (wt) malignant B-cells consisted of mixed population of tubular and smaller mitochondria compared to typical bacillus-shaped mitochondria in normal B-cells. Outer membrane was fragmented & a significant lack of the cristae structure within the mitochondria of ATM-/- DLBCL cells. (B) OCR traces from DLBCL cell line HLY. Cells were chemically inhibited for ATM signaling with KU-55933. Inhibition of ATM decreased respiration rate in HLY.

Fig. 3: The hyperpolarized ¹³C signal at ambient temperature after dissolution. Blue and red spectra correspond to metabolic process in the presence of HLY NT (+/+) and HLY SH (-/-). Both spectra are normalized with respect to maximum pyruvate signal. HLY sh-ATM (-/-) shows higher conversion of pyruvate into lactate. In the spectra pyruvate, lactate, pyruvate hydrate, alanine and bicarbonate signals are at 172.8, 184.9, 181, 178.4, and 162.7 ppm, respectively.

a: Imaging study protocol showing the temporal sequence of the applied modalities and injected tracers.

b: Axial anatomical T₂-weighted ¹H-RARE-image of a mouse bearing a subcutaneous GBM2-tumor (ROI encircled by white line) and Gd‑doped [1-¹³C]lactate-phantom (white arrow) covered by gel.

c: ADC-map showing the fitted diffusion coefficients based on 16 b-value images.

d: Mean pH-map weighted by compartment-intensities in the corresponding voxel overlaid with anatomical image.

a: Axial [1-¹³C]lactate intensity image integrated over the range displayed in c overlaid with an anatomical image.

b: Axial [1-¹³C]pyruvate intensity image integrated over the range displayed in c overlaid with an anatomical image.

c: Signal time curves for [1-¹³C]pyruvate (blue) and [1-¹³C]lactate (red) with a 3D tumor ROI. Integration ranges for intensity images in a and b are displayed with dashed lines.