Deuterium-labeled probes for diagnosing and monitoring AD using magnetic resonance deuterium imaging application techniques

Abstract

Deuterium labeled probes are provided for diagnosing and monitoring AD using magnetic resonance deuterium imaging application techniques. The invention provides application of a deuterium-labeled molecular probe or a composition containing the deuterium-labeled molecular probe to preparation of a kit or a system, and the kit or the system is used for precisely diagnosing Alzheimer's disease or dynamically monitoring the progress of the Alzheimer's disease in a subject. The deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acid, and the Alzheimer's disease is early, middle or late Alzheimer's disease. According to the invention, early accurate detection of the Alzheimer's disease is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic resonance imaging applications, and in particular to the use of deuterium-labeled probes to diagnose and monitor AD using magnetic resonance deuterium imaging application technology. Background Art

[0002] Alzheimer's disease (AD) is a neurodegenerative disease with a high mortality and disability rate. Due to its long course, serious harm and large number of patients, AD is one of the diseases with the heaviest socioeconomic burden worldwide. Currently, there is still no cure for AD, and early detection and intervention are the most effective means to slow its progression clinically.

[0003] Clinically, the diagnosis of AD mainly relies on cognitive scale tests and neuroimaging examinations. However, structural abnormalities such as brain atrophy and functional cognitive impairment appear late, and cognitive assessments are highly subjective and inaccurate, resulting in a large number of AD patients being diagnosed late or not diagnosed at all. As typical pathological characteristics of AD, molecular markers such as amyloid-β (Aβ) and neurofibrillary tangles (tau protein) are recommended for early diagnosis of clinical AD. Aβ and tau protein can be detected by cerebrospinal fluid (CSF) puncture or PET imaging. However, in 2021, the International Alzheimer's Disease Working Group wrote in Lancet Neurology that there is no necessary connection between abnormal molecular markers such as Aβ and functional cognitive impairment and AD.

[0004] More and more evidence shows that as early as 20 years before AD patients develop clinical symptoms, abnormalities in their brain molecular metabolic activity will be revealed. Pathological changes caused by abnormal metabolic pathways of energy substances such as glucose, acetate and amino acids in the brain and metabolic flux characteristics such as reaction rate and clearance rate may be the decisive factors driving the progression of AD. For example, in the early stages of the disease, the patient's brain suffers oxidative damage due to glucose glycolysis and tricarboxylic acid (TCA) cycle metabolic pathways, and the decline in the level of aerobic glycolysis in neuronal cells promotes excessive accumulation of Aβ protein and leads to excessive phosphorylation of tau protein, ultimately leading to synaptic dysfunction and impaired neurocognitive function; in addition, changes in metabolic flux characteristics such as acetate consumption rate of microglia and neurotransmitter circulation rate in neuronal cells accompanied by neuroinflammation are also considered to be the main causes of AD disease progression. JAMA magazine wrote in 2020 that energy substance metabolism represented by glucose will enter the center of the stage of AD research. Therefore, it has become an inevitable trend to deeply study the pathogenesis of AD from the perspective of brain energy substance metabolic pathways and metabolic flux characteristics, achieve early diagnosis and reflect disease progression.

[0005] However, the current cutting-edge research on AD molecular metabolism is mainly based on animal models, and related methods are difficult to achieve clinical transformation. The most fundamental reason is that the means to reflect the metabolic pathways and metabolic flows of energy substances in the human brain are extremely limited. Currently, PET and magnetic resonance imaging (MRI) molecular metabolic imaging methods that can only be used for humans can only partially reflect brain metabolic activities and cannot accurately characterize the metabolic pathways and metabolic flow characteristics of brain energy substances: (1) PET uses radioactive probes to bind to specific receptors, transporters or enzymes, and the amount of probe uptake reflects the concentration of specific markers in various brain regions. However, PET probes do not directly participate in the molecular metabolic process of the brain, nor can they detect the downstream metabolites of the probes. Therefore, they can only indirectly reflect the molecular metabolic activities of the brain; (2) Magnetic resonance imaging can directly analyze the concentrations of various neurometabolites in brain regions through spectral technology, but this method can only describe the metabolic state of the brain at a certain moment, lacks specific reflection of specific metabolic pathways, and traditional methods based on hydrogen protons ( 1 Magnetic resonance spectroscopy imaging of H) has problems such as low quantitative accuracy of metabolites, poor stability, and extremely difficult to promote clinical use. Therefore, the development of new methods that can reflect the metabolic pathways and metabolic flows of energy substances in the human brain has become an urgent need for early diagnosis of AD.

[0006] Glucose is one of the most important sources of energy for the human body. Its metabolic process in the human body is very complex, involving multiple organs and multiple biochemical pathways to generate energy. Abnormal glucose metabolism can cause a variety of diseases, the most common of which is diabetes. However, multiple evidences show that the occurrence and development of many other major diseases in the human body, such as malignant tumors, are inseparable from abnormal glucose metabolism.

[0007] Energy metabolism represented by glucose has entered the center of the research stage of various major diseases. Therefore, it has become an inevitable trend to conduct in-depth research on the pathogenesis of major diseases from the perspective of glucose metabolic pathways and metabolic flow characteristics, and to achieve early diagnosis to reflect the progression of the disease. Summary of the invention

[0008] The inventors found that in normal cells, after glucose is taken up by the cells, it will be converted into pyruvate and transported to the mitochondria to participate in the tricarboxylic acid cycle; in tumor cells, even in the presence of sufficient oxygen, tumor cells have a higher glucose uptake rate than normal cells, and are more inclined to supply energy to the cells through the glycolysis pathway, resulting in the production of a large amount of lactic acid, which is called the Warburg effect.

[0009] The inventors also found that as early as the very early stage of AD, the patient's brain suffered oxidative damage due to glucose glycolysis and tricarboxylic acid cycle metabolic pathways, and the decrease in the level of aerobic glycolysis in neuronal cells promoted the excessive accumulation of β-amyloid protein (Aβ) and caused excessive phosphorylation of tau protein, ultimately leading to synaptic dysfunction and impaired neurocognitive function.

[0010] Traditional cognitive scales and structural imaging can only be used to detect patients with mid- to late-stage AD. Detection methods based on molecular markers such as Aβ and tau protein can detect early abnormalities, but they do not have the ability to predict disease progression and have limited clinical value. In view of the defects of existing methods, the present invention develops a magnetic resonance deuterium (2H) imaging method to explore the pathogenesis of AD, achieve early diagnosis and reflect disease progression from the perspective of the metabolic pathways and metabolic flow characteristics of human brain energy substances, which is expected to fill the gap in the current AD disease diagnosis lack of abnormal information on brain energy substance metabolic activities, and provide a new idea for clarifying the pathogenesis of AD from the perspective of molecular metabolic abnormalities, achieving early diagnosis of AD and discovering new targets for AD treatment.

[0011] Accurate detection of 2H-labeled probes and their downstream metabolites is the premise and basis for visualizing the metabolic pathways and quantitative metabolic flows of energy substances in the human brain. However, the concentration of 2H-labeled metabolites in the human body is extremely low (3 to 4 orders of magnitude lower than the 1H concentration), and changes dynamically with metabolic activity, and there are differences in the metabolic characteristics of different regions of the brain. Therefore, extremely high requirements are placed on the sensitivity, temporal resolution, and spatial resolution of 2H detection. In terms of imaging principles, detection sensitivity and high temporal and spatial resolution are mutually constrained, and it is difficult for traditional methods to take all three into account at the same time. The present application realizes a method innovation of high temporal and spatial resolution detection of low-concentration 2H-labeled metabolites and precise quantification of metabolic flows.

[0012] In one aspect, provided is the use of a deuterium-labeled molecular probe or a composition comprising the deuterium-labeled molecular probe in the preparation of a kit or a system for accurately diagnosing Alzheimer's disease or dynamically monitoring the progression of Alzheimer's disease in a subject, wherein the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acid.

[0013] In one embodiment, the deuterium-labeled molecular probe is in oral or injectable form.

[0014] In one embodiment, the subject is a human or a mammal.

[0015] In one embodiment, the Alzheimer's disease is early, middle or late Alzheimer's disease.

[0016] In one embodiment, the composition comprises a first reagent or a first device for detecting metabolites of the deuterium-labeled molecular probe in the brain, preferably a magnetic resonance imaging system.

[0017] In one embodiment, the deuterium-labeled glucose is 1, 2, 3, 4, 5 or 6 deuterated glucose. In one embodiment, the deuterium-labeled glucose is [2,3,4,6,6'- 2 H 5 ]-glucose. In one embodiment, the deuterium-labeled acetate is 1, 2, 3 or 4 deuterated acetate, preferably [ 2 H 3 ]-acetate. In one embodiment, the deuterium-labeled amino acid is [ 2 In one embodiment, the acetate is selected from sodium acetate and potassium acetate.

[0018] In one embodiment, the metabolite is selected from water, glutamate or glutamate, glutamine and lactate.

[0019] In one embodiment, the composition includes a second reagent or a second device for detecting the deuterium-labeled molecular probe, preferably a magnetic resonance imaging system.

[0020] In one embodiment, the composition comprises a magnetic resonance 2 H signal acquisition component. In one embodiment, 2 In the H signal acquisition component, 1 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; 2 The number of H excitation coil channels shall not be less than 2, and the number of H receiving coil channels shall not be less than 4.

[0021] In one embodiment, the composition includes a memory storing instructions. The instructions can be executed by a processor to perform magnetic resonance 2 An imaging sequence with a shorter repetition time (TR) of H-wave spectroscopy imaging, such as a steady-state free precession free induction decay signal sequence or a balanced steady-state free precession sequence. The instructions can be executed by a processor to perform magnetic resonance imaging. 2 Quantitative estimation method of H signal.

[0022] In another aspect, there is provided an apparatus for accurately diagnosing Alzheimer's disease, comprising:

[0023] A detection unit configured to perform magnetic resonance imaging of the brain of a subject at one or more time points before and / or after administration of the substrate. 2 H imaging;

[0024] A collecting unit configured to collect magnetic resonance images of a subject's brain 2 H imaging information;

[0025] A computing unit configured to calculate a substrate consumption rate and a metabolite generation rate, and compare the calculations with those of a normal subject, wherein if the detected substrate consumption rate is reduced and / or the metabolite generation rate is increased compared to the normal subject, the detected subject suffers from or is at risk of suffering from Alzheimer's disease; wherein the substrate is a deuterium-labeled molecular probe.

[0026] The device may also include a storage unit storing instructions that, when executed by the processor, perform the magnetic resonance 2 An imaging sequence with a shorter repetition time (TR) of H-wave spectroscopy imaging, such as a steady-state free precession free induction decay signal sequence or a balanced steady-state free precession sequence, and / or the instructions when executed by a processor perform magnetic resonance imaging. 2 Quantitative calculation method of H signal.

[0027] The detection unit may include a magnetic resonance 2 H signal acquisition component. In one embodiment, 2 In the H signal acquisition component, 1 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; 2 The number of H excitation coil channels shall not be less than 2, and the number of H receiving coil channels shall not be less than 4.

[0028] In one embodiment, the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acids. In one embodiment, the subject is a human or a mammal. In one embodiment, Alzheimer's disease is early, middle or late Alzheimer's disease.

[0029] In one embodiment, the deuterium-labeled glucose is 1, 2, 3, 4, 5 or 6 deuterated glucose, preferably [2,3,4,6,6'- 2 H 5 ]-glucose. In one embodiment, the deuterium-labeled acetate is 1, 2, 3 or 4 deuterated acetate, preferably [ 2 H 3 ]-acetate. In one embodiment, the deuterium-labeled amino acid is [ 2 In one embodiment, the acetate is selected from sodium acetate and potassium acetate.

[0030] In one embodiment, the metabolite is selected from water, glutamate or glutamate, glutamine and lactate.

[0031] In one embodiment, the detection unit is a magnetic resonance imaging system, preferably a 3-10 T magnetic resonance imaging system, such as a 4, 5, 6, 7, 8 or 9 T magnetic resonance imaging system.

[0032] In one embodiment, the substrate consumption rate and metabolite production rate of a normal subject.

[0033] In yet another aspect, a method for constructing a model is provided, the model being a model for determining the progression of Alzheimer's disease, the method comprising utilizing magnetic resonance deuterium spectroscopy imaging, at one or more time points, tracking the dynamic process of metabolites of deuterium-labeled probes in different brain regions of a subject, wherein the subject has applied the deuterium-labeled molecular probe, the deuterium-labeled molecular probe being selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acids. In one embodiment, the subject is a human or a mammal, preferably a macaque. In one embodiment, the subject is a subject of early, mid-term or late Alzheimer's disease, a normal subject of different ages, such as a subject of natural aging. In one embodiment, the metabolites are selected from water, glutamate or glutamate, glutamine and lactic acid. In one embodiment, the dynamic process comprises the dynamic process of the consumption rate of the deuterium-labeled probe and the generation rate of the metabolites.

[0034] In one embodiment, the method further comprises classifying the subject's disease progression into different stages by gait measurement, cerebrospinal fluid biomarkers, PET / MRI imaging; and combining magnetic resonance imaging 2 H metabolic flux imaging, comparing metabolic flux characteristics at different stages of Alzheimer's disease, and establishing a metabolic flux characteristic map. In one embodiment, the cerebrospinal fluid biomarker is selected from Aβ and tau protein.

[0035] In yet another aspect, there is provided a device for dynamically monitoring the progression of Alzheimer's disease in a subject, comprising:

[0036] A detection unit configured to perform magnetic resonance imaging of the brain of a subject at one or more time points before and / or after administration of the substrate. 2 H imaging;

[0037] A collecting unit configured to collect magnetic resonance images of a subject's brain 2 H imaging information;

[0038] A calculation unit configured to calculate the consumption rate of the substrate and the generation rate of the metabolite, and compare them with the model described herein, wherein if the detected consumption rate of the substrate and the generation rate of the metabolite are consistent with the consumption rate of the substrate and the generation rate of the metabolite of the specific disease stage in the model, it is considered that the subject is in the specific disease stage;

[0039] The subject is administered the deuterium-labeled molecular probe, and the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acid.

[0040] The detection unit may include a magnetic resonance 2 H signal acquisition component. In one embodiment, 2 In the H signal acquisition component, 1 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; 2 The number of H excitation coil channels shall not be less than 2, and the number of H receiving coil channels shall not be less than 4.

[0041] The device may also include a storage unit storing instructions that, when executed by the processor, perform the magnetic resonance 2 An imaging sequence with a shorter repetition time (TR) of H-wave spectroscopy imaging, such as a steady-state free precession free induction decay signal sequence or a balanced steady-state free precession sequence, and / or the instructions when executed by a processor perform magnetic resonance imaging. 2 Quantitative calculation method of H signal.

[0042] In one embodiment, the subject is a human or a mammal, preferably a macaque. In one embodiment, the subject is a subject with early, middle or late stage Alzheimer's disease. In one embodiment, the metabolite is selected from water, glutamate or glutamate, glutamine and lactic acid.

[0043] In this context, deuterium-labeled glucose may be 1, 2, 3, 4, 5 or 6 deuterated glucose, preferably [2,3,4,6,6'- 2 H 5 ]-glucose such as [2,3,4,6,6'- 2 H 5 ]-D-glucose. In this context, the deuterium-labeled acetate may be 1, 2, 3 or 4 deuterated acetate, preferably [ 2 H 3 ]-acetate. Deuterium-labeled amino acids can be [ 2 H]-amino acid. The acetate may be selected from sodium acetate and potassium acetate.

[0044] The advantages of the present invention include at least:

[0045] (1) Develop methods to visualize the glucose metabolism pathways in the human brain and quantify metabolic fluxes based on magnetic resonance molecular imaging technology, and use this technology to guide the early diagnosis of AD and tumor-related diseases;

[0046] (2) For individuals, the present invention establishes data on sugar metabolism distribution and metabolic flux for individuals, guides rational diet and exercise, enhances physical fitness, and reduces the incidence of AD.

[0047] (3) Traditional cognitive scales and structural imaging can only be used to detect patients with mid- to late-stage AD. Detection methods based on molecular markers such as Aβ and tau protein can detect early abnormalities, but they do not have the ability to predict disease progression and have limited clinical value. In view of the defects of existing methods, the present invention develops a magnetic resonance deuterium (2H) imaging method to explore the pathogenesis of AD, achieve early diagnosis and reflect disease progression from the perspective of the metabolic pathways and metabolic flow characteristics of human brain energy substances, which is expected to fill the gap in the current AD disease diagnosis lack of abnormal information on brain energy substance metabolism activities, and provide a new idea for clarifying the pathogenesis of AD from the perspective of molecular metabolic abnormalities, achieving early diagnosis of AD and discovering new targets for AD treatment.

[0048] (4) Accurate detection of 2H-labeled probes and their downstream metabolites is the premise and basis for visualizing the metabolic pathways and quantitative metabolic flows of energy substances in the human brain. However, the concentration of 2H-labeled metabolites in the human body is extremely low (3 to 4 orders of magnitude lower than the 1H concentration), and changes dynamically with metabolic activities. In addition, there are differences in the metabolic characteristics of different brain regions. Therefore, extremely high requirements are placed on the sensitivity, temporal resolution, and spatial resolution of 2H detection. In terms of imaging principles, detection sensitivity and high temporal and spatial resolution are mutually restricted, and it is difficult for traditional methods to take all three into account at the same time. The present invention will collaboratively innovate in signal acquisition components, signal acquisition methods, and signal reconstruction methods, starting from the physical principles of magnetic resonance imaging and cutting-edge technologies such as artificial intelligence, to break the mutual constraints of detection sensitivity, temporal resolution, and spatial resolution, and achieve high temporal and spatial resolution detection of low-concentration 2H-labeled metabolites and method innovation for precise quantification of metabolic flows. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 :Different stages of AD disease progression and corresponding detection methods. The methods that can be used for human detection are limited to PET and magnetic resonance imaging, but neither of them can visualize the metabolic pathways of energy substances in the human brain and quantify the characteristics of metabolic flow; mass spectrometry, biochemistry, optical imaging and other methods cannot be directly applied to the human body. Magnetic resonance 2H human metabolic imaging is expected to fill this gap.

[0050] Figure 2 : Schematic diagram of magnetic resonance deuterium (2H) metabolic imaging (taking glucose metabolism as an example). The 18F-glucose used in PET molecular imaging does not participate in cell metabolism; 2H-glucose can participate in cell metabolism, and the 2H label enters the metabolite with the chemical reaction, so magnetic resonance 2H imaging can realize the visualization of metabolic pathways and the quantification of metabolic flux.

[0051] Figure 3 : Technology roadmap for the development of a 2H / 1H dual-core multi-channel coil for the human body.

[0052] Figure 4: Imaging sequences and sampling trajectories to be used in magnetic resonance 2H imaging.

[0053] Figure 5 :A framework for estimating metabolite signals and metabolic flux characteristics based on artificial intelligence and physical models

[0054] Figure 6 : Typical metabolic pathways involving glucose and acetate. * indicates that the 2H label will be transferred to downstream metabolites along with the molecular metabolic activity. The visualization of metabolic activities and quantification of metabolic flux are achieved through the quantification of downstream metabolites. The dotted line in the figure shows the pathways with impaired metabolic activity and function in the AD brain. The emergency damage of the metabolic pathway explains the phenomenon of decreased glucose consumption rate in AD patients.

[0055] Figure 7 :Using animal models to establish metabolic flux signature maps of AD disease progression.

[0056] Figure 8 :The basic framework of the feature pyramid model based on Contextual Transformer (CoT) Block.

[0057] Fig. 9 : Representative deuterium spectra of high-resolution 1H images of the hypothalamic region of 5xFAD and C57 mice 80 minutes after infusion. 5xFAD mice had higher levels of Glx in this region than C57 mice. Fig. 9 (A) is an anatomical image of the mouse brain. (B) 2H spectra of the pixel marked in (A) at 80 minutes in 5xFAD mice and C57 mice. The signal intensities of HDO, glucose (Glc), and Glx are shown in Figure 2. Fig.10 Signal intensity was averaged across all pixels in the slice containing the largest brain volume. Higher Glx production was also observed in 5xFAD, especially after 80 min. Average Glc intensity was higher in 5xFAD mice throughout the experiment, but the rate of consumption was similar in both groups.

[0058] Fig.10 : Average signal intensity of HDO, Glc and Glx over time. * indicates that there is a statistically significant difference between the two groups, p < 0.05.

[0059] Fig.11 :exist Fig. 9 Statistical plot of the signal intensity of water, glucose, and Glx at 80 min in the 6 pixels annotated in panel A. Comparison shows that 5xFAD mice have higher levels of Glx production, especially in two regions of the hypothalamus. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] It has become an inevitable trend to deeply study the pathogenesis of AD from the perspective of brain energy substance metabolic pathways and metabolic flow characteristics, achieve early diagnosis and reflect disease progression. However, PET and magnetic resonance imaging, which are currently clinically used to detect human molecular metabolism, cannot reflect abnormal brain metabolic activity from the level of metabolic pathways and metabolic flow. In response to this challenge, the present invention proposes an innovative idea of ​​using magnetic resonance deuterium (2H) spectral imaging to visualize brain energy substance metabolic pathways and quantify metabolic flow, in an effort to solve the major defect of the lack of abnormal brain molecular metabolic activity information in the current AD disease diagnosis, achieve early and accurate diagnosis of AD disease, and provide a new idea for clarifying the pathogenesis of AD from the perspective of molecular metabolic abnormalities and discovering new targets for AD treatment.

[0062] Energy metabolism is the basis of important functions such as normal brain function and antioxidant defense. Neurons are very sensitive to energy fluctuations. Impaired glucose metabolism can disrupt energy homeostasis, leading to neuronal dysfunction and cognitive impairment. At the same time, Aβ-induced neurotoxicity is also related to insufficient neuronal energy. However, behind these correlations, the specific process of abnormal energy metabolism, including the pathological characteristics of metabolic pathways such as glycolysis and TCA cycle, the catalytic and balancing effects of pyruvate and lactate, and their changing trends with the progression of AD disease are still unclear. This application is based on high-temporal and spatial resolution magnetic resonance deuterium imaging methods to explore the changes in metabolic concentrations and metabolic kinetic characteristics of metabolites such as lactate and glutamate in the early brain of AD, construct specific metabolic pathological models of glucose, acetate and amino acids and AD disease development pathways, and reveal the unique role of energy metabolism in the pathogenesis, diagnosis and intervention of AD.

[0063] in human body 2 The concentration of H-labeled metabolites is extremely low (3 to 4 orders of magnitude lower than that of 1H), and changes dynamically with metabolic activity, and there are differences in the metabolic characteristics of different brain regions, so extremely high requirements are placed on the sensitivity, temporal resolution, and spatial resolution of 2H detection. However, in terms of imaging principles, detection sensitivity and high temporal and spatial resolution are mutually constrained, and it is difficult for traditional methods to take all three into account at the same time. The present invention requires collaborative innovation in signal acquisition components, signal acquisition methods, and signal reconstruction methods to break the mutually constrained relationship between detection sensitivity, temporal resolution, and spatial resolution, and achieve high temporal and spatial resolution detection of low-concentration 2H-labeled metabolites and precise quantification of metabolic fluxes.

[0064] Method of the present invention

[0065] (1) High spatiotemporal resolution magnetic resonance based on artificial intelligence and physical modeling 2 H imaging method

[0066] This application solves the problem of high spatiotemporal resolution detection of low-concentration 2H-labeled metabolites from three aspects: magnetic resonance signal acquisition components, spectral acquisition methods, and signal reconstruction. Based on the electromagnetic field simulation method, a 2H (metabolic information) / 1H (structural information) dual-core multi-channel signal acquisition component is developed to achieve uniform and highly sensitive acquisition of dual-core signals. According to the characteristics of magnetic resonance spectroscopy and its time-varying characteristics, a high spatiotemporal resolution spectral imaging sequence is developed, and a specific sampling trajectory is designed to improve sampling efficiency. Based on physical models such as 2H magnetic resonance spectroscopy and brain metabolic dynamics, a metabolic flow parameter estimation framework driven by a physical model based on artificial intelligence is constructed, ultimately achieving accurate estimation of brain metabolic flow characteristics.

[0067] (2) Research on the relationship between AD brain energy metabolism characteristics and AD disease progression and diagnostic model

[0068] This application is based on the developed high-temporal and spatial resolution magnetic resonance deuterium (2H) imaging method and the new multi-target 2H labeled probe, which realizes the visualization of the metabolic pathways of brain glucose, amino acids, acetate and other substances and the quantitative analysis of metabolic flows such as reaction rate and clearance rate in large animal AD models, explores the metabolic flow characteristics of the brain under normal and AD pathological conditions, and constructs a metabolic flow characteristic map; through longitudinal studies, the characteristics and paths of metabolic flow pathological changes in different development stages of AD are characterized, revealing the relationship between the dynamic characteristics of brain metabolic flow and AD disease progression. Based on a convolutional neural network model with a connection attention mechanism, multi-level metabolic features for early diagnosis of AD are learned to establish an AD diagnostic model. Through the metabolic imaging of non-human primates and the establishment of a metabolic flow pathology model, the migration from animals to human applications is explored, and finally based on a 5.0T ultra-high field human magnetic resonance imaging system, the clinical application exploration of the imaging technology of the present invention and the deuterium 2H probe is carried out.

[0069] Research methods and technical routes of the present invention

[0070] High spatiotemporal resolution magnetic resonance 2H imaging method based on artificial intelligence and physical model driving

[0071] 1) Magnetic resonance 2H signal acquisition component: The performance of the magnetic resonance signal acquisition component determines the signal-to-noise ratio and detection sensitivity of the magnetic resonance signal. It is necessary to design a component that can simultaneously acquire 1H (structural information) and 2H (metabolic information) signals. However, due to the interaction between different nuclear channels in the signal acquisition component, the mutual coupling between channels, and the complex interaction between electromagnetic waves and biological tissues in high-intensity electromagnetic fields, the data transmission efficiency and receiving sensitivity will be seriously reduced.

[0072] In response to the above challenges, the present invention designs and manufactures a multi-channel frequency coil structure that can support 2H / 1H dual-core imaging based on the radio frequency electromagnetic field simulation method, and comprehensively utilizes a variety of methods such as low input impedance and low noise preamplifiers and inductive current decoupling to reduce mutual interference between channels and achieve uniform excitation and high-sensitivity acquisition of 2H / 1H dual-core signals. This application will be based on a human 5T magnetic resonance imaging system: the number of 1H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; the number of 2H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 4.

[0073] 2) Magnetic resonance 2H spectroscopy acquisition method: The magnetic resonance spectroscopy acquisition method determines the spatiotemporal resolution of metabolite detection. The ultimate goal of the present invention is to achieve quantification of metabolic flow characteristics in different brain regions, and therefore places extremely high demands on the spatiotemporal resolution of 2H signal detection. In order to achieve high spatial resolution detection, the data acquisition speed will inevitably slow down, resulting in insufficient temporal resolution, making it difficult to accurately observe the dynamic change characteristics of metabolites. In response to this challenge, the present invention develops a high spatiotemporal resolution spectral imaging sequence, designs a specific sampling trajectory based on the characteristics of magnetic resonance spectroscopy and its temporal variation characteristics, reduces unnecessary sampling points, improves sampling efficiency, and provides a methodological basis for achieving high spatiotemporal resolution deuterium (2H) imaging.

[0074] High spatiotemporal resolution 2H magnetic resonance spectrum acquisition requires joint optimization from three aspects: sequence design, sampling trajectory and undersampling scheme. Compared with traditional 1H magnetic resonance imaging, 2H has the characteristic of short longitudinal relaxation time. Based on this characteristic, imaging sequences with shorter repetition time (TR) can be selected for 2H magnetic resonance spectral imaging, such as steady-state free precession free induction decay signal (SSFP-FID) sequence or balanced steady-state free precession (balanced SSFP, bSSFP) sequence; in addition, 2H magnetic resonance sampling requires high spectral resolution, but not high requirements for spectral frequency range. According to this characteristic, corresponding sampling trajectories such as Cartesian, concentric ring or rosette can be selected; in addition, according to the characteristic of the small number of channels of 2H radio frequency coil, the undersampling scheme will mainly focus on the time dimension; the project will select the optimal acquisition sequence and corresponding sampling trajectory and undersampling scheme according to indicators such as acquisition time, signal-to-noise ratio, spectral resolution and other indicators in different application scenarios.

[0075] 3) Magnetic resonance 2H signal quantitative estimation method: The magnetic resonance signal quantitative estimation method needs to ensure the accuracy of metabolite detection and metabolic flux quantification. The present invention uses exogenous 2H labeled probes to dynamically detect downstream metabolites of the probe, thereby reflecting the characteristics of molecular metabolic activity in various brain regions. However, the 2H probe concentration within the human safety range is extremely low, resulting in 2H-labeled metabolite detection being easily interfered by background noise, making accurate detection extremely challenging.

[0076] To address this problem, the present invention will construct a metabolic flux parameter estimation framework based on artificial intelligence based on physical models such as the probe's 2H magnetic resonance spectrum and brain metabolic dynamics to achieve accurate estimation of brain metabolic flux characteristics. After the 2H-labeled probe is taken into the body, the probe and its downstream metabolites can be detected on the magnetic resonance deuterium (2H) spectrum, and its signal model can be expressed as:

[0077]

[0078] Among them, S is the overall spectral signal intensity obtained by reconstruction, Amm(xx,TT) is the content of metabolite mm at spatial position xx and time TT, M is the number of metabolites, and LLmm is the distribution characteristics of metabolites in the frequency domain. Directly solving the above equation will lead to large fitting errors due to the low signal-to-noise ratio, and the content of each metabolite will be inaccurately estimated.

[0079] In the above model, since the probes and their metabolites are known, the distribution characteristics Lm of each metabolite in the frequency domain can be learned in advance using the collected spectral data:

[0080]

[0081] Among them, a n,m is the weighting coefficient, The number of spectral sampling points is 2Ns, the echo time is TE, and the sampling interval is Δf. The chemical shift of the metabolite is fm, and the main magnetic field offset in the actual sampling is Δfm. The chemical shift fm of each metabolite can be obtained based on prior knowledge. Based on the actual collected spectral data and the above physical model, the distribution characteristics Lm of each metabolite in the frequency domain are learned in advance.

[0082] In addition to the frequency domain characteristics of metabolites, glucose and its downstream metabolites also meet pharmacokinetic constraints, and this prior information will also be used to build the network model. The simplified deuterated glucose kinetic model is described in the literature (Neuroimage Clin. 2022; 33: 102932), and its downstream products include lactate, glutamate + glutamine and water. The concentration changes of glucose and downstream products are affected by factors such as reaction rate and clearance rate:

[0083]

[0084] Where [Glc], [Lac] and [Glx] are the concentrations of glucose, lactate and glutamate + glutamine, respectively (mM); kg is the transfer rate of glucose between blood and brain tissue (min -1 ); is the concentration of glucose in plasma (mM); f is the enrichment of deuterium; v is the extracellular volume fraction outside the blood vessels; Vmax = Vlac + Vglx, that is, the glucose consumption rate (Vmax) is the sum of the lactate and glutamate + glutamine synthesis rates (Vlac and Vglx); km is the constant for glucose absorption; klac and kglx are the clearance rates of lactate and glutamate + glutamine, respectively.

[0085] Study on the association between brain metabolic flux characteristics and AD disease progression and diagnostic model

[0086] 1) Construction of normal and AD pathological models in non-human primates

[0087] At present, the animal models used for AD research are mainly transgenic mice. However, the mouse brain tissue is very different from the human brain in terms of structure, function and genetic background, which makes it difficult to directly apply the research results found in mice to humans. Non-human primates are highly similar to humans in terms of genetic background, brain structure, pathological characteristics and aging process, and are ideal experimental animals in AD research. The present invention uses naturally aged monkeys with AD pathological characteristics and behavioral symptoms as experimental subjects.

[0088] The present invention uses deuterium (2H) labeled probes (including [2,3,4,6,6'-2H5]-glucose, [2H3]-acetate and [2H]-amino acids, etc.) and uses magnetic resonance deuterium (2H) spectroscopy imaging to track the dynamic metabolic process of deuterium (2H) labeled probes in different brain regions of animal models. Taking the metabolic pathways of glucose and acetate as an example (e.g. Figure 6 As shown in the figure, through magnetic resonance 2H spectral imaging, the concentration information of 2H-labeled downstream metabolites such as lactate, glutamate and water in different brain regions at different times can be quantified, so as to realize the visualization of brain energy substance metabolic pathways and quantitative analysis of metabolic flows, thereby constructing and comparing the characteristics and differences of energy substance intake and metabolic activities of the brain under normal and AD disease states.

[0089] 2) Metabolic flow characteristic map of AD disease progression

[0090] Studying and establishing the metabolic characteristics of AD at different stages of disease progression is of great significance for the diagnosis and treatment effect evaluation of AD. The present invention will explore the metabolic flow characteristics corresponding to different stages of AD progression. Macaques of different ages (middle-aged: about 10 years old, elderly about 20 years old) were selected, and the AD disease progression was divided into different stages through traditional gait measurement, cerebrospinal fluid biomarkers (Aβ, tau protein), and PET / MRI imaging (Alzheimers Dement, 2011; 7: 270–279). Using the above animal model, combined with magnetic resonance 2H metabolic flow imaging, the metabolic flow characteristics of different stages of AD were compared, and a metabolic flow characteristic map was established, such as Figure 7 3) Research on AD diagnostic model based on CoT feature pyramid deep learning Based on magnetic resonance 2H metabolic signals and dynamic feature information, the present invention constructs a feature pyramid model based on Contextual Transformer (CoT) Block to learn multi-level features in AD magnetic resonance 2H metabolic flow feature images and 1H structural images.

[0091] The model mainly consists of an encoder, a decoder, a jump connection, and a feature combination. In the encoder part, a series of feature maps of different resolutions are extracted, followed by normalization, activation functions, and a maximum pooling layer; in the decoder part, each step includes upsampling the feature map, then halving the number of feature channels, concatenating with the corresponding feature map in the encoder, and finally merging the resulting feature maps to produce classification probability values, such as Figure 8 shown.

[0092] The encoder and decoder parts introduce a new attention module CoT module, which makes full use of contextual information to guide the learning of dynamic attention matrix, thereby enhancing the ability of image feature representation. The CoT module integrates contextual information mining and self-attention learning into a unified architecture. It first uses k×k×k groups of convolutions for all adjacent key values ​​to spatially associate contexts with each key value. The learned key value with contextual information is represented by K1. Two consecutive 1×1×1 convolutions are also used in this module to obtain the relationship matrix A. The two convolutions are represented by Wθ and Wδ respectively. Before this, K1 must be connected with Q to obtain the relationship matrix A=[K 1 , Q]W θ W δ . This process no longer forms an isolated QK pair, but a relational mapping that can obtain rich contextual feature information. Next, the relational matrix A is multiplied by V to obtain K2. K2 is the dynamic context representation of the input, which can capture the dynamic feature interaction between the inputs. Finally, K1 and K2 are fused through the attention mechanism to obtain the output result.

[0093] Example

[0094] The following examples are provided to illustrate the present invention. It should be understood by those skilled in the art that the examples are merely illustrative and not limiting. The present invention is limited only by the scope of the appended claims.

[0095] Embodiment 1:

[0096] method

[0097] Mouse AD model 5xFAD (n=3, male, 6 months old) was used and compared with normal mice C57 (n=3, male, 6 months old). T2-weighted high-resolution anatomical images were obtained on a 9.4T animal system (uMR 9.4T, United Imaging LifeScience Instruments, Wuhan, China) using a dedicated mouse brain coil with parameters of TR / TE=3000ms / 35.56ms, Matrix=370*370, FOV=17mm*17mm, and slice thickness=0.2mm.

[0098] Deuterium magnetic resonance imaging was performed using a Bruker 11.7T MR system and a homemade 2H / 1H coil (Du F, Yuan J, LiN, et al. Deuterium metabolism imaging of rat brain at 9.4T using a double-nuclear transceiver [C] / / Proc. Intl. Soc. Mag. Reson. Med. 2022, 3243). During imaging acquisition, all mice were anesthetized with isoflurane. The dose was dynamically adjusted in the range of 0.5-1.8% to keep the respiratory rate at 30-40 times per minute throughout the experiment. All mice were injected with [2, 3, 4, 6, 6'- 2 H 5 ]-D-glucose (Dingbang Biotechnology Co., Ltd., Shenzhen, China) was dissolved in normal saline at a dose of 3 g / kg. T2-weighted anatomical images were also acquired at 11.7 T, but due to the coil 1 H imaging performance was impaired and image quality was reduced. Chemical shift imaging (CSI) sequence was used to detect deuterium-labeled substrates in the brain. Imaging parameters were: TR = 150 ms, matrix = 8x8x8, field of view = 16 mm x16 mm x 16 mm, bandwidth = 2 kHz, acquisition points = 256, average = 150, acquisition time = 16.5 min. The imaging sequence was repeated 7 times, 1 time before glucose infusion as a baseline and 6 times after infusion.

[0099] The whole brain spectral data were exported and MATLAB was used to write code for Fourier inverse transform, line broadening and phase correction. The spectrum was then fitted to the mixed Lorentzian model using the least squares fitting algorithm (ZouC, Ruan Y, Li H, et al. A new deuterium-labeled compound [2,3,4,6,6'- 2 H 5 ]-D-glucosefor deuterium magnetic resonance metabolic imaging[J].NMR in Biomedicine,2023,36(7):e4890). The signal intensity was normalized relative to the HDO obtained before glucose infusion. 1 The H image is registered with the image at 9.4T to determine 2 Anatomical localization of H spectral data.

[0100] result

[0101] Fig. 9 Representative deuterium spectra of high-resolution 1H images of the hypothalamus region of 5xFAD and C57 mice at 80 minutes after infusion. The level of Glx in this region was higher in 5xFAD mice than in C57 mice. Fig. 9 (A) is an anatomical image of a mouse brain. (B) 2H spectra of the labeled pixel 1 in (A) at 80 minutes in 5xFAD mice and C57 mice.

[0102] The signal intensities of HDO, glucose (Glc) and Glx are shown in Figure 2 Fig.10 Signal intensity was averaged across all pixels in the slice containing the largest brain volume. Higher Glx production was also observed in 5xFAD, especially after 80 min. Average Glc intensity was higher in 5xFAD mice throughout the experiment, but the rate of consumption was similar in both groups.

[0103] Fig.10 is the average signal intensity of HDO, Glc and Glx over time. * indicates that there is a statistically significant difference between the two groups, p < 0.05.

[0104] Fig.11 Shown in Fig. 9 Statistical plot of the signal intensity of water, glucose, and Glx at 80 min in the 6 pixels annotated in panel A. Comparison shows that 5xFAD mice have higher levels of Glx production, especially in two regions of the hypothalamus.

[0105] The present invention found that in the AD model, the rate of Glx production at the whole brain level was higher. However, in this study, this significant difference was observed as early as 6 months of age (rather than 14 months of age). In addition, 3D CSI found that the rate of Glx production in the hypothalamus region of 5xFAD mice was always higher. In addition, there was no difference in glucose consumption between the two groups. This observation may be different from the 18F-FDG-PET study (Bouter C, Henniges P, Franke TN, et al. 18F-FDG-PET detects drastic changes in brain metabolism in the Tg4-42 model of Alzheimer's disease [J]. Frontiers in aging neuroscience, 2019, 10: 425). However, in another study using 13C-labeled glucose in a 7-month-old 3xTgAD mouse model, it was found that the flux of 13C-labeled glucose was even higher than that of normal mice (Sancheti H, Patil I, Kanamori K, et al. Hypermetabolic state in the 7-month-old triple transgenic mouse model of Alzheimer's disease and the effect of lipoic acid: a 13C-NMR study [J]. Journal of Cerebral Blood Flow & Metabolism, 2014, 34 (11): 1749-1760). The inconsistent findings will be further investigated. The comparison will be fully extended to PET and even the behavioral characteristics of animal models. In addition, a larger group will continue to complete this preliminary study.

[0106] Impaired early brain energy metabolism is a driving factor in the progression of Alzheimer's disease (AD). This example explores the potential of using deuterium magnetic resonance imaging to characterize glucose metabolism in AD mouse models. 2 H 5]-D-glucose was used to perform deuterium magnetic resonance spectroscopy imaging on 5xFAD and C57 mice. Preliminary results showed that 5xFAD mice had a higher production rate of glutamine / glutamate (Glx) than normal mice, especially in the hypothalamic region. The work of the present invention may provide inspiration for the development of a new method for early detection of AD based on abnormal glucose metabolism. Alzheimer's disease (AD) is a neurodegenerative disease with a high mortality and disability rate. Early diagnosis is the most effective way to control the development of the disease, but it remains a huge challenge (Anand R, Gill KD, Mahdi AA. Therapeutics of Alzheimer's disease: Past, present and future [J]. Neuropharmacology, 2014, 76: 27-50). Recent studies have found that brain energy metabolism disorders occur many years before the onset of clinical symptoms and play an important role in disease progression (Jagust W. Imaging the evolution and pathophysiology of Alzheimer disease [J]. Nature Reviews Neuroscience, 2018, 19 (11): 687-700). In this study, we explored the potential of using deuterium metabolic imaging to reveal the characteristics of AD glucose metabolism in animal models (De Feyter HM, Behar KL, Corbin ZA, et al. Deuterium metabolic imaging (DMI) for MRI-based 3D mapping of metabolism in vivo [J]. Science advances, 2018, 4 (8): eaat7314).

[0107] Although the present invention has been described at some length and in some detail with respect to several described embodiments, it is not intended to be limited to any such detail or embodiment or any specific embodiment, but should be interpreted with reference to the appended claims to provide the widest possible interpretation of such claims in light of the prior art, and thus effectively cover the intended scope of the present invention. It should be understood that although the present invention has been described in conjunction with its detailed description, the foregoing description is intended to illustrate rather than limit the scope of the present invention, which is limited by the scope of the appended claims. Other aspects, advantages and modifications are within the scope of the appended claims.

Claims

1. Use of a deuterium-labeled molecular probe or a composition comprising a deuterium-labeled molecular probe in the preparation of a kit or system for accurately diagnosing Alzheimer's disease in a subject or dynamically monitoring the progression of Alzheimer's disease, wherein the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acids, and preferably the deuterium-labeled molecular probe is in oral or injectable form; preferably, the subject is a human or a mammal; preferably, Alzheimer's disease is early, middle or late Alzheimer's disease.

2. The use according to claim 1, wherein the composition comprises a first reagent or a first device for detecting metabolites of the deuterium-labeled molecular probe in the brain, preferably a magnetic resonance imaging system.

3. The use according to claim 1 or 2, wherein the deuterium-labeled glucose is 1, 2, 3, 4, 5 or 6 deuterated glucose, preferably [2, 3, 4, 6, 6'- 2 H5]-glucose, such as [2,3,4,6,6'- 2 H5]-D-glucose; the deuterium-labeled acetate is 1, 2, 3 or 4 deuterated acetate, preferably [ 2 H3]-acetate; the deuterium-labeled amino acid is [ 2 H]-amino acid; preferably, the acetate is selected from sodium acetate and potassium acetate.

4. The use according to claim 2 or 3, wherein the metabolite is selected from water, glutamic acid or glutamate, glutamine and lactic acid.

5. The use according to any one of claims 1 to 4, wherein the composition comprises a second reagent or a second device for detecting the deuterium-labeled molecular probe, preferably a magnetic resonance imaging system; Preferably, the composition comprises a magnetic resonance 2 H signal acquisition component, preferably, 2 In the H signal acquisition component, 1 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; 2 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 4; The composition includes a memory storing instructions which, when executed by a processor, perform high temporal and spatial resolution magnetic resonance 2 H spectrum acquisition, including sequence design, sampling trajectory, and undersampling process; Preferably, sequence design includes magnetic resonance 2 Imaging sequences with shorter repetition time (TR) of H-spectroscopy imaging, such as those based on steady-state free precession free induction decay signal sequences or balanced steady-state free precession sequences; Preferably, the sampling trajectory is selected from the sampling trajectory of Cartesian, concentric ring or rosette and / or The undersampling process is an undersampling scheme focused on the time dimension; and / or The instructions, when executed by a processor, perform magnetic resonance 2 Quantitative estimation method of H signal.

6. A device for accurately diagnosing Alzheimer's disease, comprising: A detection unit configured to perform magnetic resonance imaging of the brain of a subject at one or more time points before and / or after administration of the substrate. 2 H imaging; A collecting unit configured to collect magnetic resonance images of a subject's brain 2 H imaging information; preferably, the collection unit includes a magnetic resonance 2 H signal acquisition component, preferably, 2 In the H signal acquisition component, 1 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; 2 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 4; a calculation unit configured to calculate a consumption rate of a substrate and a generation rate of a metabolite, and compare the calculations with the substrate consumption rate and the metabolite generation rate of a normal subject, wherein if the detected substrate consumption rate is decreased and / or the metabolite generation rate is increased compared to the normal subject, the detected subject suffers from or is at risk of suffering from Alzheimer's disease; The substrate is a deuterium-labeled molecular probe; Optionally, a memory unit storing instructions that, when executed by the processor, perform high temporal and spatial resolution magnetic resonance 2 H spectrum acquisition, including sequence design, sampling trajectory, and undersampling process; Preferably, sequence design includes magnetic resonance 2 Imaging sequences with shorter repetition time (TR) of H-spectroscopy imaging, such as those based on steady-state free precession free induction decay signal sequences or balanced steady-state free precession sequences; Preferably, the sampling trajectory is selected from the sampling trajectory of Cartesian, concentric ring or rosette and / or The undersampling process is an undersampling scheme focused on the time dimension; and / or The instructions, when executed by a processor, perform magnetic resonance 2 Quantitative estimation method of H signal.

7. The device according to claim 6, wherein the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acid, preferably, the subject is a human or a mammal; preferably, Alzheimer's disease is early, middle or late Alzheimer's disease.

8. The device according to claim 7, wherein the deuterium-labeled glucose is 1, 2, 3, 4, 5 or 6 deuterated glucose, preferably [2,3,4,6,6'- 2 H5]-glucose such as [2,3,4,6,6'- 2 H5]-D-glucose; the deuterium-labeled acetate is 1, 2, 3 or 4 deuterated acetate, preferably [ 2 H3]-acetate; the deuterium-labeled amino acid is [ 2 H]-amino acid; preferably, the acetate is selected from sodium acetate and potassium acetate; Preferably, the metabolite is selected from water, glutamate or glutamate salt, glutamine and lactic acid.

9. The apparatus according to any one of claims 6 to 8, wherein the detection unit is a magnetic resonance imaging system, preferably a 3-10 T magnetic resonance imaging system, such as a 4, 5, 6, 7, 8 or 9 T magnetic resonance imaging system.

10. The device according to any one of claims 6 to 9, wherein the substrate consumption rate and metabolite production rate are those of a normal subject.

11. A method for constructing a model, wherein the model is a model for determining the disease progression of Alzheimer's disease, the method comprising tracking the dynamic process of metabolites of a deuterium-labeled probe in different brain regions of a subject at one or more time points using magnetic resonance deuterium spectroscopy imaging, wherein the subject is administered the deuterium-labeled molecular probe, and the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acids, preferably, the subject is a human or a mammal, preferably a macaque; preferably, the subject is a subject with early, middle or late stage Alzheimer's disease, a normal subject of different ages, such as a naturally aged subject; preferably, wherein the metabolite is selected from water, glutamate or glutamate, glutamine and lactate; preferably, the dynamic process comprises the dynamic process of the consumption rate of the deuterium-labeled probe and the generation rate of the metabolite; Preferably, the method comprises magnetic resonance 2 H spectroscopy imaging with short repetition time (TR) imaging sequences, such as those based on steady-state free precession free induction decay signal sequences or balanced steady-state free precession sequences, and / or magnetic resonance imaging 2 Quantitative estimation method of H signal.

12. The method of claim 11, further comprising classifying the subject's disease progression into different stages by gait measurement, cerebrospinal fluid biomarkers, PET / MRI imaging; and combining magnetic resonance imaging with 2 H metabolic flux imaging, comparing the metabolic flux characteristics of different stages of Alzheimer's disease and establishing a metabolic flux characteristic map; preferably, the cerebrospinal fluid biomarkers are selected from Aβ and tau proteins.

13. A device for dynamically monitoring the progression of Alzheimer's disease in a subject, comprising: A detection unit configured to perform magnetic resonance imaging of the brain of a subject at one or more time points before and / or after administration of the substrate. 2 H imaging; A collecting unit configured to collect magnetic resonance images of a subject's brain 2 H imaging information; preferably, the collection unit includes a magnetic resonance 2 H signal acquisition component, preferably, 2 In the H signal acquisition component, 1 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 2; 2 The number of H excitation coil channels is not less than 2, and the number of receiving coil channels is not less than 4; A calculation unit configured to calculate the consumption rate of the substrate and the generation rate of the metabolite, and compare them with the model described in claim 11 or 12, wherein if the detected consumption rate of the substrate and the generation rate of the metabolite are consistent with the consumption rate of the substrate and the generation rate of the metabolite of the specific disease stage in the model, it is considered that the subject is in the specific disease stage; Optionally, a memory unit storing instructions that, when executed by the processor, perform high temporal and spatial resolution magnetic resonance 2 H spectrum acquisition, including sequence design, sampling trajectory, and undersampling process; wherein the subject is administered the deuterium-labeled molecular probe, the deuterium-labeled molecular probe is selected from deuterium-labeled glucose, deuterium-labeled acetate or deuterium-labeled amino acid, preferably, the subject is a human or a mammal, preferably a macaque; preferably, the subject is a subject with early, middle or late stage Alzheimer's disease; preferably, the metabolite is selected from water, glutamate or glutamate, glutamine and lactate; Preferably, sequence design includes magnetic resonance 2 Imaging sequences with shorter repetition time (TR) of H-spectroscopy imaging, such as those based on steady-state free precession free induction decay signal sequences or balanced steady-state free precession sequences; Preferably, the sampling trajectory is selected from the sampling trajectory of Cartesian, concentric ring or rosette and / or The undersampling process is an undersampling scheme focused on the time dimension; and / or The instructions, when executed by a processor, perform magnetic resonance 2 Quantitative estimation method of H signal; Preferably, the deuterium-labeled glucose is 1, 2, 3, 4, 5 or 6 deuterated glucose, preferably [2,3,4,6,6'- 2 H5]-glucose such as [2,3,4,6,6'- 2 H5]-D-glucose; the deuterium-labeled acetate is 1, 2, 3 or 4 deuterated acetate, preferably [ 2 H3]-acetate; the deuterium-labeled amino acid is [ 2 H]-amino acid; preferably, the acetate is selected from sodium acetate and potassium acetate.

Images