Deuterium Magnetic Resonance Imaging Using Deuterated Water-Induced ²H-Tissue Labeling Allows Monitoring Cancer Treatment at Clinical Field Strength

Accurate and noninvasive assessment of tumor response following radiation and/or chemotherapy is essential for managing therapeutic regimens in patients. Traditionally, anatomic imaging techniques, such as CT and MRI have been used to plan radiotherapy and assess tumor response. In addition, morphologic evaluation by MRI is now an indispensable technique for accurately diagnosing diseases, such as cancer and tracking therapeutic effects. However, anatomic imaging techniques are insufficient for the early detection of therapeutic responses after therapy because a tumor's size may not change for weeks or months after treatment.

Molecular imaging visualizes and monitors metabolic processes in the living body at the molecular level. Diagnostic methods, such as PET and MRI that target molecular processes, have been clinically applied in a wide range of conditions, and expectations for use in personalized medicine and effective drug discovery have been increasing. Although PET can track the biochemical changes in tumors after therapy, it is frequently postponed because of the confounding effects of inflammation and occasionally low predictive values. Moreover, PET-CT scans have relatively poor spatial resolution and their frequent use may expose patients to excessive ionizing radiation (1).

MRI can also be used to visualize bodily structures, but it has limited molecular-level imaging capabilities because of its low detection sensitivity for trace amounts of substances. The sensitivity of MRI depends on the polarization rate of the nuclear spin. At normal temperatures, the directions of nuclear spins are complicated, and the signals cancel each other out. Therefore, it is necessary to increase nuclear spin polarization to increase MRI sensitivity. Since its development more than a decade ago (2), hyperpolarization via dynamic nuclear polarization (DNP) has rapidly become a well-established technique to enhance the NMR signals of low-gamma nuclei that have a long liquid-state longitudinal relaxation time (T₁), such as ¹³C, ¹⁵N, ⁶Li, and ¹²⁹Xe (3–6). Hyperpolarization MRI enables metabolic imaging of administered probes, so it is used to visualize the microenvironment of the living body by observing differences in the metabolism of molecules in diseased areas, such as those in cancer. For example, when cancer is tested by DNP–magnetic resonance spectroscopy (MRS), pyruvic acid, a major metabolic intermediate of glucose, is labeled with ¹³C, hyperpolarized, and administered intravenously as a metabolic probe. By measuring real-time metabolism by MRI, the difference in metabolism between normal cells and cancerous cells can be evaluated (6). Currently, DNP is in practical use and clinical research, but it is laborious to disseminate the DNP device to general hospitals because the introductory cost is high, and its technical requirements restrict its applications to a limited number of suitable metabolites.

In recent years, deuterium metabolic imaging (DMI) has been demonstrated, and recent MR studies have concluded that deuterated metabolites possess desirable properties for the noninvasive mapping of biochemical processes of healthy and diseased organs (7). A promising metabolic imaging approach is to track the fate of orally administered deuterated glucose because it is taken up and metabolized into different products in tissues, such as the heart, brain, pancreas, liver, and cancerous tissue (8–12).

The stable hydrogen isotope, ²H, is used as a tracer in metabolic imaging for studying the dynamic metabolism of compounds. D₂O was used in research as a ²H uptake label by administration of D₂O to mice and as a free diffusivity tracer for blood flow and tissue circulation (13). Deuterium MR measurements are of growing interest in metabolic imaging because ²H has an integer spin of one, which causes a quadrupolar magnetic moment, and it is sensitive to molecular orientation and produces faster relaxation than ¹H (14, 15). When compared with ¹H, ²H has a lower natural abundance (∼0.015%) and a gyromagnetic ratio (6.54 MHz/T). These characteristics cause a reduced MR signal. However, the short longitudinal relaxation time (T₁) enables rapid signal averaging with an SNR per unit acquisition time gain that is approximately 3×–10× higher for ²H than for ¹H. This has led to the use of ²H-labeled substrates as in vivo metabolic tracers and to an increased interest in deuterium MR measurements (9, 16). Therefore, DMI has become an emerging MR modality for investigating energy metabolism in vivo (9, 12, 17). MR measurements that depend on tracing ²H are featured by its insensitivity to magnetic field inhomogeneity, increased sensitivity through signal averaging, and technical simplicity. Moreover, ²H is a stable nonradioactive isotope that does not require special handling facilities. Accordingly, ²H can be traced at clinical magnetic field strengths without the need for special techniques or facilities. In addition, D₂O administration is known to lead to the incorporation of ²H into metabolic pathways and tissue labeling. Tracing the ²H-labeled organs by MRI should also help characterize D₂O kinetics in normal and diseased body tissues.

The study aim was to investigate D₂O kinetics in bodily tissues and track ²H-labeled tissues during oral D₂O administration using a magnetic field strength of 1.5T, which is widely used in clinical practice. We also hoped to demonstrate the feasibility of tumor imaging and delineation of the efficacy of anticancer radiation and chemotherapeutics in a human pancreatic cancer model based on a 1.5T deuterium MRI.

Heavy water (D₂O, 98.8%) and deuterated methanol (Methanol-D4, 99.95%) were obtained from Sigma-Aldrich Corp. Pyruvic acid (1–¹³C, 99%) was obtained from Cambridge Isotope Laboratories, Inc. Trityl radical OX063 was supplied by GE Healthcare. Bevacizumab (trade name, Avastin) was bought from Chugai Pharmaceutical Co., Ltd. Gemcitabine hydrochloride was bought from Tokyo Chemical Industry Co., Ltd. All other chemicals were of reagent-grade quality and obtained from commercial sources.

All animal care and experimental protocols were approved by the Committee for Animal Research and Welfare of Medical School, Gifu University and conducted according to the recommendations of the Committee for Care and Use of Laboratory Animals, Gifu University. Female 4- to 6-week-old BALB/c-nu mice (RRID:IMSR_CRL:194) weighing 15–20 g were purchased from Charles River Laboratories Japan, Inc. Before the experiments, all animals were acclimated to a normal diet for 1 week with free access to distilled water and appropriate food (MF diet, Oriental Yeast Co.). Every effort was made to minimize the number of mice used and their suffering.

MIA PaCa-2 human pancreatic carcinoma cells (catalog no. CRL-1420TM, RRID: CVCL_0428) were purchased from ATCC and C26 murine colorectal carcinoma cells (Cell# RCB2657, RRID:CVCL_0240) were obtained from RIKEN Cell Bank. Cells were grown in DMEM (Wako Pure Chemical Industries, Ltd.), supplemented with 10% heat-inactivated FBS (MP Bio LLC.) and penicillin–streptomycin (100 U/mL penicillin and 100 mg/mL streptomycin, Gibco Life Technologies) at 37°C in a humidified incubator containing 5% CO₂. A pathologic tumor model was developed by subcutaneous administration of 0.8–1.2 × 10⁶ mouse cells into the right leg of each mouse. The tumor diameters were measured with a Vernier caliper. When the tumor volume reached approximately 500 mm³, the experiments were started. The following approximate formula was used to determine the tumor volume: tumor volume = length/2 × width × height. Before the experiments, the body weight was between 18g and 20g.

To evaluate the feasibility of visualizing ²H at 1.5T MRI, a phantom of nine wells was designed and prepared by a 3D printer for use. Different concentrations of D₂O in H₂O [10%–90% (v/v)] were added to each well. ¹H-specific and ²H-specific coils at resonance frequencies of 64 and 9.8 MHz, respectively, were used to scan the phantom. The phantom was imaged by proton and deuterium MRI methods at different repetition times (¹H: 500–3,000 ms, ²H: 100–500 ms).

For the in vivo imaging studies, mice were provided free access to drinking water containing 30% v/v D₂O during imaging and treatment periods. The control mice were allowed free access to regular tap water (0.015% D₂O) instead of deuterated water.

The effects of X-ray irradiation, bevacizumab, and gemcitabine on tumor progression and the ²H MR signals in tumor tissues were determined in mice randomly divided into four groups (n = 4): the control group (intraperitoneal injection of 100 μL physiologic saline, at 1-day intervals), bevacizumab group (intraperitoneal injection of bevacizumab 10 mg/kg, at 1-day intervals), gemcitabine group (intraperitoneal injection of gemcitabine 120 mg/kg at 1-day interval), and the irradiation group (single dose of 20 Gy X-ray irradiation). For irradiation, the mice were restrained without anesthesia in a custom jig. The xenograft-bearing legs were then irradiated in an X-ray medical linear accelerator (LINAC), Primus (Siemens Healthcare). During all the treatment experiments, mice were allowed free access to deuterated water simultaneously with treatment. Mice were sacrificed by cervical dislocation and xenografts were removed, fixed in 10% formalin for histopathologic examination, or kept at −80°C for the NMR and hyperpolarized ¹³C-MRS studies.

Xenograft-bearing leg images were obtained on a 1.5T permanent magnet MRI system (Medalist) obtained from Japan Redox Ltd. Mice were anesthetized with isoflurane (3% for induction and 1.0%–2.0% for maintenance in medical air, 350 mL/minute). Using a PicoM nonmagnetic rectal fiber optic temperature sensor (Opsens Inc.), the mouse body temperature was monitored and kept at 36°C ± 1°C using a warm air flow set at 38°C. The mouse body was then fixed in a prone position on a customized holder using adhesive skin tape. The holder was set into a ¹H or ²H volume coil of the MR scanner.

Deuterium MR scanning was performed using gradient echo coronal or transverse 2D deuterium imaging. After a quick assessment of the sample position by a fast low-angle shot pilot sequence, ²H images were obtained for the head and forelimbs, abdomen, and hind limbs. During the treatment, images of xenograft-bearing legs were also acquired. One 50-mm–thick coronal or 15-mm–thick transverse slice was acquired at a 100 ms repetition time (TR), echo time (TE) of 10 ms, 64 accumulations, and a 64 × 64 matrix. During each scan, a marker phantom, which was a 5-mm diameter vial containing 5% D₂O in H₂O (v/v), was placed next to the target ROI as a ²H reference signal. Coronal and transverse T₁- and T₂-weighted proton images were also acquired by a spin echo sequence with TR of 500 ms or 3,000 ms, TE of 10 ms or 40 ms, two or four accumulations, five 2-mm slices, and a 64 × 64 matrix.

Hyperpolarization was achieved by using 22 μL of 11 M ¹³C-pyruvate with 15 mmol/L OX063. A HyperSense DNP polarizer (Oxford Instruments) was set at a microwave irradiation frequency of 93.42 GHz at 100 mW and 1.4K for 45 minutes. The hyperpolarized probe solution was mixed with 3 mL dissolution buffer containing 50 mmol/L trisaminomethane, 0.3 mmol/L EDTA, and 60 mmol/L NaOH. Immediately after dissolution, 250 μL of hyperpolarized probe solution was mixed in a prewarmed NMR tube containing 350 μL of 1:1 tumor homogenate in PBS and 50 μL of 500 mmol/L nicotinamide adenine dinucleotide. ¹³C spectral acquisition was then done for 260 seconds using a 1.4T Spinsolve 60 Carbon High-Performance Benchtop NMR instrument (Magritek) at a flip angle = 10° and TR = 4 seconds.

Tumor tissues were collected and homogenized in 100 mmol/L phosphate buffer at a 1:1 rate and centrifuged at 5,000 rpm for 20 minutes at 4°C. Supernatants were collected, and methanol-d₄ was added at a final concentration of 100 mmol/L in a total of 600 μL. Samples containing the reference standard were placed into a 5 mm NMR tube and examined for ²H content by ²H-NMR. For that purpose, an Avance 800 NMR spectrometer (Bruker BioSpin) was used. The spectrometer was operated at a ¹H frequency of 800.15 MHz and a 5-mm ¹H inverse detection probe with triple-axis gradient coils was used for all measurements. The ²H spectra were successively acquired at 25°C, with 4,096 complex points covering 1,476 Hz for ²H. NMR data processing and analysis were performed using the TopSpin NMR software package v3.6.2 (Bruker BioSpin).

After collection, xenografts were fixed in 10% formalin and embedded in paraffin. Tissues were cut into 3 μm–thick sections, deparaffinized, and stained with hematoxylin and eosin (H&E) or used for IHC analysis. For antigen retrieval, the sections were placed in citrate buffer (pH 6.0) and autoclaved at 110°C for 10 minutes. The sections were then washed three times in PBS for 5 minutes each and then washed and blocked with 3% hydrogen peroxide in methanol for 10 minutes to remove endogenous peroxidase. Next, the sections were rinsed three times in PBS for 5 minutes again. Nonspecific binding sites were blocked by adding PBS (pH 7.4) containing 2% BSA (Fujifilm Holdings Corporation) for 40 minutes at room temperature and then incubated overnight at 4°C with primary antibodies against the cellular proliferation marker, anti-human Ki-67 (dilution, 1:100; M7240; Dako, RRID:AB_2142367), tumor angiogenesis marker, and anti-mouse CD31 (dilution, 1:50; DIA 310; Dinova Gmbh, RRID:AB_2631039). Primary antibodies were detected by using Simple Stain Mouse MAX PO (rat; Nichirei Bioscience) for 60 minutes at room temperature. The bound complex was visualized using SIGMAFAST 3,30-diaminobenzidine tablets (Sigma-Aldrich Corporation) and counterstained with hematoxylin. The sections were mounted and images were obtained by using a Keyence all-in-one fluorescence BZ-X800 microscope (KEYENCE) and analyzed with imaging cellSens software (RRID:SCR_014551, Olympus).

All experiments were performed at least three times. All results shown in the graphs are expressed as the mean ± SD. Differences in the group means were determined by a two-tailed Student t test. The minimum level of significance was set at P < 0.05.

Representative data of MRI, histopathology, and hyperpolarized ¹³C-MRS experiments are presented in this article and the Supplementary Data. Raw NMR data can be accessed at https://zenodo.org/record/8333098. All other materials, data, and protocols generated in this study are available upon reasonable request from the corresponding author.

We first used a phantom containing various concentrations of D₂O in H₂O (v/v) to evaluate the sensitivity of ²H detection by a 1.5T MR scanner and ²H-specific coil. The D₂O was clearly visible at concentrations as low as 10% (Fig. 1A and B). The signal intensity at 9.8 MHz was directly related to the D₂O content. Acceptable image quality of both H₂O (by proton MRI) and D₂O by deuterium MRI were obtained at a repetition time of 500 ms in both cases (Fig. 1B).

To evaluate in vivo deuterium MRI at 1.5T, mice with MIA PaCa-2 or C26 xenografts were freely allowed to drink 30% D₂O and imaged for D₂O kinetics. During the 10-day administration, the distribution and accumulation of ²H in different mouse tissues were studied. Imaging of each of the three body sections was performed. The distribution of ²H in the tissues was observable as early as 1 day after D₂O administration (Fig. 2A; Supplementary Fig. S1A). Moreover, upon continuous drinking of D₂O, the ²H atoms of D₂O gradually replaced the protons in the tissues to various degrees, and the signal strength was gradually increasing on the 3rd, 7th, and 10th days (Fig. 2A; Supplementary Fig. S1A). Proton images of the whole mouse body were also acquired, and each image was superimposed to obtain a clear image of the whole-body ²H distribution (Fig. 2B and C; Supplementary Fig. S1B and S1C). Interestingly, the highest ²H tissue accumulation was observed in the tumor tissues (Fig. 2D; Supplementary Fig. S1D).

We studied the feasibility of applying deuterium MRI for monitoring the efficacy of the different anticancer interferences. The MIA PaCa-2 xenograft-bearing legs of mice were subjected to a single dose of radiation, and the mice were freely allowed to drink 30% D₂O. Deuterium MRI of the tumor was then performed on days 1, 3, and 10. The changes in the tumor size were also measured. The accumulation of ²H in the tumor tissue was significantly reduced in the treated mice at days 1, 3, and 10 posttreatments (Fig. 3A and B). On the other hand, there were no significant differences in the tumor volumes during this early stage of tumor treatment. The effect of treatment on tumor volumes was later confirmed by continuous monitoring of tumor volume and by proton imaging up to day 27. The MRI data also showed a significant decrease in the tumor volume in the treated mice xenografts (Fig. 3C and D).

On a similar basis, the effect of anticancer drugs on ²H MR signal was assessed in the xenograft-bearing mice after treatment with bevacizumab or gemcitabine. The mice were intraperitoneally injected while allowed to drink D₂O for 1 week, and deuterium MRI was performed on days 1, 3, and 7 of treatment. Similar to the pattern for radiotherapy, despite no obvious change in tumor volume, the ²H MR signals in the tumor xenografts were significantly lower during the treatment period in the treated mice than in the control mice (Fig. 4A–D). As measured by total ingested volumes per day, there was no significant difference in the amount of D₂O consumed by the treated mice groups.

To quantitatively confirm the observed reductions in the ²H content during treatment, we used the ²H-NMR technique to measure the relative ²H content in the extracts of equal volumes of tumor tissues from the treated and control mice. The NMR spectrometer was equipped with a specific probe for ²H acquisition that provided signals relevant to the concentrations of ²H in the samples. As shown in Supplementary Fig. S2, the ²H content was significantly lower in the gemcitabine-treated mice than in the controls.

To observe the effect of anticancer treatments, an ex vivo study of hyperpolarized ¹³C-MRS was conducted to investigate the changes in the Warburg effect in the control and treated mice. Tumor homogenates were mixed with hyperpolarized ¹³C-pyruvate, and the ¹³C-metabolite signals were measured by NMR in real time immediately after mixing. In a pattern matching that of the deuterium MR data, the calculated lactate-to-pyruvate ratios were significantly lower in all treated groups than in the controls (Fig. 5A).

When the tumor samples were examined histopathologically, the H&E-stained sections revealed significantly bigger necrotic areas in the bevacizumab- and gemcitabine-treated mice but not the irradiated mice. In addition, the IHC analysis revealed no significant differences in the cell proliferation or microvascular density markers (Fig. 5B–D).

The growing demand for metabolic-specific imaging strategies has rekindled interest in deuterium MRI, a reliable procedure that involves administering a ²H-enriched substrate (glucose, acetate, fumarate, etc.) followed by imaging to examine its distribution and metabolism. This technique enables the investigation of metabolic processes under healthy and diseased conditions. Despite its low natural abundance, the short relaxation time of ²H enables rapid radiofrequency pulses without saturation and proper image acquisition (18). Furthermore, deuterium MRI can provide insights into the spatial biochemistry of living tissues in conjunction with routine proton imaging for anatomic localization (19–21). Therefore, deuterium MRI is analogous to other techniques that depend on the administration of metabolites as probes to track metabolic processes. Moreover, deuterium MRI may be superior to these techniques. It allows tracking of the uptake and distribution of administered D₂O in target tissues. D₂O has been used to study cell kinetics, protein synthesis, and metabolism (22–24). The low natural abundance of ²H (0.015%) is a major advantage for the detection of an exogenous deuterated substrate and its products above the background and serves as a quantitative internal reference. For example, when D-[6,6′-²H₂]-glucose was administered, its metabolites of lactate, glutamine, and glutamate were visually measurable through ²H-MRS after approximately 1 hour. That made three-dimensional DMI maps of deuterated glucose metabolites using chemical-shift imaging also feasible (9). However, instead of the costly high doses of deuterated glucose, the administration of D₂O water in this study showed that it is still possible to clearly visualize the deuterated molecules’ fate in tissues. Moreover, the variations in uptake, accumulation, and distribution among these tissues are a further advantage of D₂O for studying the accumulation of newly formed molecules in highly proliferative tissues, such as tumors (21).

Previous reports have shown that ingestion of water with ²H content >40% affects the homeostasis of mouse bodies and reduces daily intake (25). Therefore, in this experiment, we allowed the mice to freely drink water containing 30% D₂O for a maximum of 10 days during which image acquisition was conducted. First, imaging was performed separately of the mouse's head and fore limbs, abdomen, and lower limbs. The distribution and accumulation of ²H in these anatomical areas were visually measurable as early as 1 day after D₂O administration. Interestingly, the deuterium MR imaging data revealed that accumulation of ²H was highest in tumor xenografts than in other tissues from the first day of administration (Fig. 2D). That accumulation further increased daily during the 10 days of the study. Similar results were also observed in the C26 xenograft-bearing mice (Supplementary Fig. S1D) which confirm that highly proliferative cancer tissues uptake ²H more than other organs. These findings could be explained by cell proliferation in the presence of ²H atoms inducing ²H-labeling of the newly formed cell molecules by replacing the carbon bonds with protium atoms, especially given that cancerous cells are characterized by rapid cell division and growth. These growing cells have a high demand for carbon and hydrogen to produce new cellular biomass and maintain high rates of proliferation (26, 27). During the synthesis of this cellular biomass, enzymatic processes can use ²H instead of ¹H as a substrate, leading to the formation of new carbon–deuterium bonds that are not exchangeable with protons. After D₂O administration, ²H is also known to be integrated into amino acids through transamination and other intermediary pathways, and new proteins are then synthesized from these amino acids (24). In addition, ²H is incorporated into the DNA of dividing cells during nucleotide synthesis through the constitutive de novo nucleotide synthesis pathway (21–23). Thus, simple oral administration of D₂O may enable in vivo labeling of highly proliferating cells, including cancerous cells (23, 28).

The ²H-labeling of cell components after administration of D₂O has long been used through the use of deuterated water or glucose for measuring cell proliferation in different tissues, including slow-turnover cells such as mammary epithelial cells, colon epithelial cells, and vascular smooth muscle cells (29). Administration of D₂O also has been used for in vivo follow-up of the proliferation rates of pancreatic islets (22) as a key factor in the pathogenesis of diabetes mellitus. The observed results of higher ²H-labeling rates in tumor tissues are in accordance with previous reports in which D₂O enrichment was used as a probe and showed increased DNA synthesis rates after the administration of a carcinogen (30). These data led to our hypothesis that hydrogen will usually accumulate in rapidly proliferating tumor cells during progression because of the high metabolic activity and involvement in the build-up of the newly formed cell biomass and DNA. The data also prompted us to hypothesize that tracing ²H signals in tumors may help monitor tumor progression and assess the irradiation efficacy early.

During treatments, tumor morphology may remain unchanged for weeks or months, rendering anatomic imaging methods inadequate for early detection of therapeutic response. In this study, we evaluated a comprehensive application of deuterium MRI for monitoring tumor development and early treatment efficacy using radiotherapy or chemotherapy, with a special focus on its potential clinical applications in tumor theranostics. Systemic D₂O labeling in four groups of mice bearing MIA PaCa-2 tumor xenografts was induced. The mice were then treated with radiation, bevacizumab, gemcitabine, or saline and examined by deuterium MRI. Bevacizumab is a recombinant mAb used as a first-line therapy for treating several types of cancers; it blocks angiogenesis by inhibiting VEGF A, a protein that stimulates angiogenesis in cancer (31–33). Gemcitabine is a pyrimidine antagonist that is widely used in chemotherapy regimens for a number of carcinomas. The drug is used as a first-line treatment for pancreatic cancer (34) by interfering with the DNA of cancer cells, inhibiting the synthesis of DNA necessary for cell division and suppressing cancer division and growth. Despite no morphologic changes, significant reductions in ²H accumulation were observed in MIA PaCa-2 tumor xenografts following ionizing radiation or the intraperitoneal injections of bevacizumab or gemcitabine anticancer drugs. This finding suggested the ability of anticancer therapy to induce cell apoptosis, disruption of blood vessels, and loss of the newly formed carbon–deuterium bonds. This ability was partially confirmed by the higher cellular apoptosis observed in the treated tumors according to histopathologic data (Fig. 5B). In addition, the ex vivo hyperpolarization MRS study of the changes in the metabolism of the hyperpolarized ¹³C-pyruvate probe in the tumor homogenate of the control and treated mice also confirmed the in vivo results (Fig. 5A). Following injection of prepolarized ¹³C-pyruvate, changes in the relative lactate observed has been shown to be a marker for tumor progression or early treatment response (35). In addition, Kaggie and colleagues recently compared ¹³C-HP-MRS using DMI as a noninvasive method for the first time and showed how they simultaneously probed metabolism (36). Their data demonstrated that the two methods provided complementary data. Although it may still be challenging to obtain high-quality metabolite images by deuterium MRI as those of ¹³C-HP-MRS, this issue might be overcome by improving the tools used. In addition, using D₂O as a target probe instead of deuterated glucose may have solved this issue partially. Our images showed acceptable to good signal-to-noise ratios obtained in a few minutes at 1.5T in mice administered D₂O orally.

The D₂O-labeling/deuterium MRI technique is characterized by its operational simplicity, and lower cost of D₂O than other deuterated molecules, such as glucose. Moreover, the fact that D₂O enrichment of proliferative tissues is measurable by deuterium MRI provides an additional advantage. The ease of visualization of D₂O enrichment of body tissues using deuterium MRI is also a great advantage for outpatient or field studies. This technique may be considered suitable for routine applications in monitoring proliferating tissues because of its simplicity, adaptability, reproducibility, and predictability.

In this study, deuterium MRI at 1.5T was successfully used to track the ²H signal distributed throughout mouse tissues during D₂O administration with higher ²H observed in both the MIA PaCa-2 and C26 tumor tissues. The feasibility of using deuterium MRI after D₂O administration to detect anticancer treatment responses in a pancreatic cancer model was demonstrated in vivo. The tested radiotherapy or chemotherapeutics decreased the ²H in the tumor tissues before macroscopic changes occur. The method is efficient, safe, economical and promising.

Deuterium MRI at 1.5T was successfully used to track the ²H signal distributed throughout mouse tissues during D₂O administration. We found that the oral uptake of D₂O in mice at 30% followed by deuterium MRI revealed a higher accumulation of ²H in the tumor tissues than in controls. In addition, the administration of radiation treatment or chemotherapeutic agents decreased the observable ²H in the tumor tissue before the occurrence of macroscopic changes in tumor morphology. Our findings suggest that assessment of tumor progression and treatment based on deuterium MRI after ingestion of D₂O, as a potential imaging biomarker, could be useful for assessing responses to radiotherapy and chemotherapy. The method should be an efficient, safe, and noninvasive approach using D₂O labeling to monitor tumor progression and treatment in vivo and is simple, economical, and applicable in animals and humans. This research indicated the potential of deuterium MRI as a clinical method for understanding and monitoring anticancer agents, particularly for the treatment of pancreatic cancer. Overall, the results indicate that deuterium MRI, a minimally invasive approach, may offer new ways of studying and characterizing tumor progression and treatment.

No disclosures were reported.

H. Asano: Conceptualization, resources, data curation, formal analysis, visualization, methodology, writing–original draft, writing–review and editing. A.E. Elhelaly: Conceptualization, resources, data curation, software, formal analysis, visualization, methodology, writing–original draft, writing–review and editing. F. Hyodo: Conceptualization, resources, supervision, funding acquisition, visualization. R. Iwasaki: Conceptualization, resources, data curation, software, formal analysis, methodology. Y. Noda: Data curation, software, methodology. H. Kato: Conceptualization, data curation, software. K. Ichihashi: Data curation, investigation, methodology. H. Tomita: Resources, data curation, software, formal analysis, supervision. M. Murata: Resources. T. Mori: Conceptualization, resources, data curation, supervision, visualization. M. Matsuo: Conceptualization, resources, supervision, funding acquisition, methodology.

This research was supported by AMED under Grant Number JP19cm0106435h0002 and JP21fk0108555h0001. This work was also supported by JSPS KAKENHI (grant numbers 22K07768 and 20KK0253), MEXT Quantum Leap Flagship Program (MEXT Q-LEAP) grant number JPMXS0120330644, and JST Fusion Oriented Research for disruptive Science and Technology (grant number JPMJFR2168).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.