Alzheimer's disease prediction method and system, electronic equipment and medium

Abstract

The invention provides an Alzheimer's disease prediction method and system, electronic equipment and a medium. The Alzheimer's disease prediction method comprises the following steps: acquiring multi-modal data of a to-be-detected object; pre-processing the multi-modal data, and automatically and quantitatively calculating an analysis index of a gap around a blood vessel based on diffusion tensor imaging so as to obtain processed data to be detected; using an Alzheimer's disease prediction model to predict the processed to-be-tested data to obtain an Alzheimer's disease risk level of the to-be-tested object; and generating a visual decision report for the to-be-detected object based on the Alzheimer's disease risk level. According to the Alzheimer's disease prediction method, the features of the multi-modal data can be automatically fused, the early Alzheimer's disease can be accurately predicted in a graded manner, and the early warning precision and accuracy are improved.

Description

Technical Field

[0001] This application belongs to the fields of medical health and artificial intelligence technology, and relates to an Alzheimer's disease prediction method, and in particular to an Alzheimer's disease prediction method, system, electronic device and medium. Background Technology

[0002] Alzheimer's disease (AD) is the most common neurodegenerative disease in the elderly, accounting for approximately 60% to 80% of dementia cases. Currently, the diagnosis of AD mainly relies on cerebrospinal fluid analysis. -Amyloid-beta (A) The detection of α and β proteins, combined with invasive detection methods such as positron emission tomography (PET) and intrathecal injection, is used. However, PET technology is expensive, and intrathecal injection carries certain risks, limiting its widespread application in large-scale early screening.

[0003] In recent years, the brain's glymphatic system has received widespread attention as an important pathway for waste removal. The Diffusion Tensor Image Analysis Along the Perivascular Space Index (DTI-ALPS), based on diffusion tensor imaging (DTI), indirectly reflects the functional status of the glymphatic system by quantifying the diffusion characteristics of water molecules within the perivascular space. Studies have shown that the DTI-ALPS index progressively decreases in AD patients, indicating that glymphatic system dysfunction is closely related to disease progression. Summary of the Invention

[0004] This application provides an Alzheimer's disease prediction method, system, electronic device, and medium to address the problems of low accuracy and inefficiency in Alzheimer's disease risk prediction in the prior art.

[0005] In a first aspect, this application provides an Alzheimer's disease prediction method. The Alzheimer's disease prediction method includes: acquiring multimodal data of a subject; preprocessing the multimodal data and automatically quantifying and calculating an analysis index along the perivascular space based on diffusion tensor imaging to obtain processed test data; using an Alzheimer's disease prediction model to predict the processed test data to obtain the Alzheimer's disease risk level of the subject; and generating a visual decision report for the subject based on the Alzheimer's disease risk level.

[0006] In this application, multimodal data of the test subjects are preprocessed, and an analysis index along the perivascular space based on diffusion tensor imaging is automatically calculated. This processed data is then input into an Alzheimer's disease prediction model for prediction, obtaining the Alzheimer's disease risk level of the test subjects. Based on the Alzheimer's disease risk level, a visualized decision report of the test subjects is generated. The Alzheimer's disease prediction method of this application can automatically fuse features of multimodal data to accurately and hierarchically predict early Alzheimer's disease, improving the accuracy and precision of early warning.

[0007] In one implementation of the first aspect, the automatic quantization calculation of the perivascular space analysis index based on diffusion tensor imaging includes: preprocessing the diffusion tensor image using principal component analysis to obtain a corrected diffusion tensor image; estimating a diffusion tensor model based on the corrected diffusion tensor image; spatially registering the diffusion tensor model using a standard space to obtain the initial coordinates of the region of interest; calculating the distribution of the projection and joint components of the initial coordinates based on the tensor to obtain the peak position of the region of interest; and calculating the perivascular space analysis index based on the peak position.

[0008] In one implementation of the first aspect, the calculation based on the peak position to obtain the perivascular space analysis index based on diffusion tensor imaging includes: generating a region of interest with a preset radius centered on the peak position; and calculating based on the region of interest to obtain the mean of the tensor in three directions as the perivascular space analysis index based on diffusion tensor imaging.

[0009] In one implementation of the first aspect, the multimodal data of the subject includes clinical data, imaging data, and radiomics data; the clinical data includes the subject's age, sex, education level, cognitive score, apolipoprotein E genotype, and plasma p-tau217 concentration; the imaging data includes the subject's hippocampal volume and perivascular space analysis index based on diffusion tensor imaging; and the radiomics data includes radiomics characteristics of the subject's hippocampal volume.

[0010] In one implementation of the first aspect, obtaining the radiomics features of the hippocampal volume of the subject includes: preparing a fast gradient echo sequence using three-dimensional T1-weighted magnetization to acquire the hippocampal volume of the subject; segmenting the hippocampal volume using a hybrid segmentation algorithm to obtain hippocampal subregions; performing intracranial volume normalization on the hippocampal subregions to obtain residual volumes; correcting the hippocampal volume of the subject using the residual volumes to obtain corrected hippocampal volumes; and extracting features from the corrected hippocampal volumes to obtain the radiomics features of the hippocampal volume of the subject.

[0011] In one implementation of the first aspect, obtaining the radiomics features of the hippocampal volume of the subject to be tested further includes: performing multiple screenings on the extracted hippocampal volume image features by combining stability screening, dimensionality reduction using the maximum correlation minimum redundancy feature selection algorithm, and biological significance verification, so as to obtain the radiomics features of the hippocampal volume of the subject to be tested.

[0012] In one implementation of the first aspect, the training method for the Alzheimer's disease prediction model includes: acquiring multimodal data of Alzheimer's patients, wherein the feature data includes clinical data, imaging data, and radiomics data; classifying and labeling the multimodal data of Alzheimer's patients to obtain multimodal data with different risk levels; preprocessing the multimodal data with different risk levels to obtain a training dataset; and training a classification network model combined with a multi-head attention mechanism using the training dataset to obtain the Alzheimer's disease prediction model.

[0013] Secondly, this application provides an Alzheimer's disease prediction system. The Alzheimer's disease prediction system includes: a data acquisition module for acquiring multimodal data of the subject; a data processing module for preprocessing the multimodal data and automatically quantifying and calculating an analysis index along the perivascular space based on diffusion tensor imaging to obtain processed test data; a model prediction module for predicting the processed test data using an Alzheimer's disease prediction model to obtain the Alzheimer's disease risk level of the subject; and a result display module for generating a visual decision report for the subject based on the risk level.

[0014] Thirdly, this application provides an electronic device. The electronic device includes: a memory for storing a computer program; and a processor for executing the computer program stored in the memory to cause the electronic device to perform the Alzheimer's disease prediction method as described in any one of the first aspects.

[0015] Fourthly, this application provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the Alzheimer's disease prediction method described in any one of the first aspects.

[0016] As described above, the Alzheimer's disease prediction method, system, electronic device, and medium described in this application have the following beneficial effects:

[0017] This application enables the automatic and accurate quantification of the perivascular space analysis index based on diffusion tensor imaging (DTI), and correlates this index with Alzheimer's disease risk prediction, achieving accurate and efficient prediction of Alzheimer's disease risk. Based on the obtained Alzheimer's disease risk level, a corresponding personalized decision report is generated, improving the early warning capability for asymptomatic Alzheimer's patients. Attached Figure Description

[0018] Figure 1 The diagram shown illustrates an application scenario of the Alzheimer's disease prediction method described in this application.

[0019] Figure 2 This diagram illustrates the structure of the mid-cloud interaction scenario in these implementation methods.

[0020] Figure 3 The diagram shown is a flowchart of the Alzheimer's disease prediction method described in the embodiments of this application.

[0021] Figure 4 The diagram shown is a schematic representation of the Alzheimer's disease prediction method described in an embodiment of this application.

[0022] Figure 5 The diagram shown is a flowchart of the Alzheimer's disease prediction method described in the embodiments of this application.

[0023] Figure 6 The diagram shown is a flowchart of the Alzheimer's disease prediction method described in the embodiments of this application.

[0024] Figure 7 The diagram shows a flowchart illustrating the training method of the Alzheimer's disease prediction model described in this application embodiment.

[0025] Figure 8 The diagram shown is a structural schematic of the Alzheimer's disease prediction system described in an embodiment of this application.

[0026] Figure 9 The diagram shown is a structural schematic of the electronic device described in an embodiment of this application.

[0027] Component designation explanation

[0028] 1 Risk prediction device 11 Data acquisition equipment 12 Local processor 13 Display terminal 2 End-to-Cloud Interaction System 20 terminal 21 cloud server 100 Alzheimer's Disease Prediction System 110 Data acquisition module 120 Data processing module 130 Model prediction module 140 Results Display Module 900 electronic devices 910 memory 920 processor 930 monitor S11~S14 step S121~S125 step S21~S25 step S31~S34 step Detailed Implementation

[0029] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0030] It should be noted that in the embodiments of this application, the words "optionally" or "for example" indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "optionally" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "optionally" or "for example" is intended to present the relevant concepts in a specific manner.

[0031] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0032] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0033] Predicting the risk of Alzheimer's disease (AD) and early intervention are major challenges in the field of neuroscience. Existing predictive models suffer from significant technical bottlenecks. First, although there is an abundance of potential biomarkers, including genetics (such as APOE), the methods remain largely undeveloped. 4) Body fluid biomarkers (such as plasma p-tau217) and multimodal neuroimaging (such as structural MRI and DTI), but most studies are still limited to single-modal data analysis, lacking an effective multimodal data fusion framework, and failing to fully reveal the synergistic and complementary effects between risk factors of different dimensions, thus limiting further improvement in prediction accuracy. Secondly, regarding imaging biomarkers, the diffusion tensor imaging-based perivascular space analysis (DTI-ALPS) index, as a way to assess brain lymphoid system function and reflect A... Novel non-invasive indicators of clearance capacity have shown great potential. However, their calculation heavily relies on manual delineation by professionals, a cumbersome process prone to subjective errors. The lack of efficient and standardized automated calculation procedures hinders their widespread adoption in large-scale population screening and routine clinical applications.

[0034] At least to address the aforementioned problems, embodiments of this application provide an Alzheimer's disease prediction method. The Alzheimer's disease prediction method includes: acquiring multimodal data of a subject; preprocessing the multimodal data and automatically quantifying and calculating an analysis index along the perivascular space based on diffusion tensor imaging to obtain processed test data; using an Alzheimer's disease prediction model to predict the processed test data to obtain the Alzheimer's disease risk level of the subject; and generating a visual decision report for the subject based on the Alzheimer's disease risk level.

[0035] In this embodiment, the multimodal data of the test subject is preprocessed, and an analysis index along the perivascular space based on diffusion tensor imaging is automatically calculated. This preprocessed data is then input into an Alzheimer's disease prediction model for prediction, obtaining the Alzheimer's disease risk level of the test subject. Based on the Alzheimer's disease risk level, a visual decision report of the test subject is generated. The Alzheimer's disease prediction method of this application can automatically fuse features from multimodal data to accurately and hierarchically predict early-stage Alzheimer's disease, improving the accuracy and precision of early warning.

[0036] Figure 1 This diagram illustrates an application scenario of the Alzheimer's disease prediction method described in this application. The risk prediction device 1 can be used to implement the Alzheimer's disease prediction method provided in this application embodiment, but the application scenarios of the Alzheimer's disease prediction method provided in this application embodiment are not limited to this. Figure 1 The risk prediction device 1 shown is as follows. Figure 1 As shown, the risk prediction device 1 includes a data acquisition device 11, a local processor 12, and a display terminal 13. The Alzheimer's disease prediction method provided in this embodiment can be applied to the local processor 12.

[0037] in, Figure 1The local processor 12 can be a single local processor, a cluster of multiple local processors, or a cloud computing center, etc., and is not specifically limited here. Although Figure 1 Only one data acquisition device 11, one local processor 12, and one display terminal 13 are shown in the diagram, but it should be understood that... Figure 1 The examples in this paper are only for understanding this solution. The specific number of local processors 12 and display terminals 13 should be flexibly determined based on the actual situation.

[0038] In some other implementations, the risk prediction device 1 may not include a display terminal 13, but only a local processor 12 with display functionality and a data acquisition device 11. The Alzheimer's disease prediction method provided in this application embodiment can be applied to the local processor 12. The local processor 12 with display functionality may include tablet computers, PDAs, mobile phones, personal computers, and voice interaction devices, or it may be a monitoring device, a face recognition device, etc., which are not limited here.

[0039] In some other implementations, the Alzheimer's disease prediction method described in this application can be applied to edge-cloud interaction scenarios. Figure 2 This diagram illustrates the structure of the endpoint-cloud interaction scenario in these implementation methods. For example... Figure 2 As shown, the terminal-cloud interaction system 2 includes a terminal 20 and a cloud server 21. The terminal 20 and the cloud server 21 can communicate with each other, and the communication method is not limited to wired or wireless.

[0040] The terminal 20 can be mobile or fixed. For example, it can be a wireless terminal or a wired terminal. A wireless terminal can refer to a device with wireless transceiver capabilities, which can be deployed indoors, outdoors, and in medical laboratories. The terminal 20 can be a mobile phone, tablet computer, laptop computer, etc., and is not limited thereto. The cloud server 21 can include one or more servers, or one or more processing nodes, or one or more virtual machines running on the server. The cloud server 21 can also be referred to as a server cluster, management platform, data processing center, etc., and is not limited thereto in this embodiment.

[0041] The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0042] The following embodiments of this application provide a method for predicting Alzheimer's disease, which, for example, can be achieved through... Figure 1 The local processor 12 shown Figure 2 The cloud server 21 shown is used to implement this. Figure 3 The diagram shown is a flowchart illustrating the Alzheimer's disease prediction method described in an embodiment of this application. Figure 3As shown, the Alzheimer's disease prediction method includes steps S11 to S14.

[0043] Step S11: Obtain multimodal data of the object under test.

[0044] Step S12 involves preprocessing the multimodal data and automatically quantizing and calculating the perivascular space analysis index based on diffusion tensor imaging to obtain the processed test data.

[0045] Step S13: Use the Alzheimer's disease prediction model to predict the processed test data to obtain the Alzheimer's disease risk level of the test subject.

[0046] Step S14: Generate a visual decision report for the subject of the test based on the risk level of Alzheimer's disease.

[0047] Among some possible implementations, Figure 4 The diagram shown is a schematic representation of the Alzheimer's disease prediction method described in an embodiment of this application. Figure 4 As shown, the multimodal data of the subjects included three categories: clinical data, imaging data, and radiomics data. The multimodal data included the subjects' age, sex, education level, cognitive score, apolipoprotein E genotype, plasma p-tau217 concentration, hippocampal volume, diffusion tensor imaging-based perivascular space analysis index, and radiomics characteristics of hippocampal volume. The multimodal data underwent preprocessing, and the diffusion tensor imaging-based perivascular space analysis index was automatically quantified to obtain the processed data. An Alzheimer's disease prediction model was used to predict the Alzheimer's disease risk level of the subjects. Specifically, the Alzheimer's disease risk level was divided into three categories: high-risk, intermediate-risk, and low-risk. High-risk refers to a risk of developing Alzheimer's disease within 2 years greater than 35%, intermediate-risk refers to a risk of developing Alzheimer's disease within 5 years of 15-25%, and low-risk refers to a risk of developing Alzheimer's disease within 5 years of less than 5%. A visual decision report is generated for each subject based on their Alzheimer's disease risk level. This report includes the probability of developing Alzheimer's disease, eigenvalue contribution, and intervention recommendations. For example, it may suggest clinical anti-A treatment for high-risk individuals. For treatment, it is recommended that individuals at intermediate risk undergo another cognitive assessment or A4 test 6 months later. -PET scans confirm whether the risk of Alzheimer's disease has changed; annual follow-up is recommended for low-risk individuals.

[0048] In this embodiment, the multimodal data of the test subject is preprocessed, and an analysis index along the perivascular space based on diffusion tensor imaging is automatically calculated. This preprocessed data is then input into an Alzheimer's disease prediction model for prediction, obtaining the Alzheimer's disease risk level of the test subject. Based on the Alzheimer's disease risk level, a visual decision report of the test subject is generated. The Alzheimer's disease prediction method of this application can automatically fuse features from multimodal data to accurately and hierarchically predict early-stage Alzheimer's disease, improving the accuracy and precision of early warning.

[0049] Figure 5 The diagram shown is a flowchart illustrating the Alzheimer's disease prediction method described in an embodiment of this application. Figure 5 As shown, step S12 includes steps S121 to S125.

[0050] Step S121: Preprocess the diffusion tensor image using principal component analysis to obtain the corrected diffusion tensor image.

[0051] Step S122: Estimate the diffusion tensor model based on the corrected diffusion tensor imaging.

[0052] Step S123: Spatial registration of the diffuse tensor model is performed using standard space to obtain the initial coordinates of the region of interest.

[0053] Step S124: Calculate the distribution of the projection and joint components of the initial coordinates based on the tensor to obtain the peak position of the region of interest.

[0054] Step S125: Calculate the perivascular space analysis index based on the peak position to obtain the index based on diffusion tensor imaging.

[0055] In some possible implementations, the Marchenko-Pastur principal component analysis method is used to preprocess the diffusion tensor imaging to eliminate Gibbs ringing artifacts, correct distortion caused by eddy currents, and implement N4 bias field correction to eliminate image intensity inhomogeneities, thus obtaining corrected diffusion tensor images. The Marchenko-Pastur principal component analysis method is a statistically based denoising method for diffusion-based magnetic resonance imaging. N4 bias field correction is an advanced algorithm for intensity inhomogeneity correction in magnetic resonance imaging.

[0056] The diffusion tensor model is estimated based on a modified diffusion tensor imaging (DTI) model. First, the region of interest (ROI) is initially located. Specifically, the positions of the RIOs (Regions of Interest) based on the projection and association fibers along the perivascular space analysis index (PAI) are pre-defined on the FA template in the standard space of the Montreal Neurological Institute (MNI). Then, spatial registration maps the RIOs to the DTI space to obtain their initial coordinates. Next, the RIOs are precisely located. Specifically, the distribution of the projection and association components of the initial coordinates is calculated based on the tensor. The projection or association diffusion components of each voxel are calculated within a cuboid near the initial coordinates of the RIOs. The maxima near the initial coordinates of the RIOs—the "peaks"—are identified to determine the peak position of the RIOs. Finding these maxima is essentially finding the local maxima of the diffusion index, where the diffusion index is the grayscale value of the FA image. Fine-tuning is then performed using the four bilateral maxima to precisely locate the RIOs on both sides of the lateral ventricle body, ensuring that the four RIOs lie on the same left-right line. The peak location was used as a reference to obtain the perivascular space analysis index based on diffusion tensor imaging.

[0057] In one embodiment of this application, the calculation based on the peak position to obtain the perivascular space analysis index based on diffusion tensor imaging includes: generating a region of interest with a preset radius centered on the peak position; and calculating based on the region of interest to obtain the mean of the tensor in three directions as the perivascular space analysis index based on diffusion tensor imaging.

[0058] In some possible implementations, four regions of interest with a radius of 3 mm are drawn using the peak location as the center, and the mean values ​​of the tensor components in the three directions within each region of interest are calculated. The formula for calculating the perivascular space analysis index based on diffusion tensor imaging is as follows:

[0059] ,

[0060] in, The mean diffusivity along the left and right directions within the region of interest. The mean diffusion value along the left-right direction within the region of interest is projected. To combine the mean diffusion in the vertical direction within the region of interest, The mean diffusion value along the front-to-back direction is projected within the region of interest. It should be noted that the coordinate system of the region of interest is the same as the coordinate system of the diffusion tensor image obtained from the scan.

[0061] In one embodiment of this application, the multimodal data of the test subject includes clinical data, imaging data, and radiomics data. Clinical data includes the test subject's age, sex, education level, cognitive score, apolipoprotein E genotype, and plasma p-tau217 concentration; imaging data includes the test subject's hippocampal volume and perivascular space analysis index based on diffusion tensor imaging; and radiomics data includes radiomics characteristics of the test subject's hippocampal volume.

[0062] In some possible implementations, cognitive scoring includes the total score of the Montreal Cognitive Assessment (MoCA) after corrected education, the total score of the Mini-Mental State Examination (MMSE), and the total score of the Clinical Dementia Rating Scale (CDR). Standardized data collection, after normalization, is input into the model. Apolipoprotein E genotype (ApoE4) is a binary variable, including carriers... 4 alleles and non-carriers There are four alleles, two types. Plasma p-tau217 concentration is the concentration of phosphorylated tau protein at site 217 in the blood.

[0063] Figure 6 The diagram shown is a flowchart illustrating the Alzheimer's disease prediction method described in an embodiment of this application. Figure 6 As shown, obtaining the radiomics features of the hippocampal volume of the object under test includes steps S21 to S25.

[0064] Step S21: Prepare a fast gradient echo sequence using three-dimensional T1 weighted magnetization to acquire the hippocampal volume of the object under test.

[0065] Step S22: The hippocampal volume is segmented using a hybrid segmentation algorithm to obtain hippocampal subregions.

[0066] Step S23: Standardize the intracranial volume of the hippocampal subregion to obtain the residual volume.

[0067] Step S24: The residual volume is used to correct the hippocampal volume of the test object to obtain the corrected hippocampal volume.

[0068] Step S25: Feature extraction is performed on the corrected hippocampal volume to obtain the radiomics features of the hippocampal volume of the subject under test.

[0069] In some possible implementations, a fast gradient echo sequence is prepared using 3D-T1WI MPRAGE with three-dimensional T1-weighted magnetization to acquire the hippocampal volume of the subject. A hybrid segmentation algorithm is used to segment the hippocampal volume, accurately dividing it into 12 hippocampal subregions. Residual volumes are obtained through intracranial volume normalization. These residual volumes are then used to correct the hippocampal volume of the subject, yielding the corrected hippocampal volume. Finally, radiomics features of the subject's hippocampal volume are obtained through multi-scale omics feature extraction and a three-stage feature selection process.

[0070] In one embodiment of this application, obtaining the radiomics features of the hippocampal volume of the subject to be tested further includes: performing multiple screenings on the extracted hippocampal volume image features by combining stability screening, dimensionality reduction using the maximum correlation minimum redundancy feature selection algorithm, and biological significance verification, so as to obtain the radiomics features of the hippocampal volume of the subject to be tested.

[0071] In some possible implementations, radiomics features of several acquired hippocampal volumes are screened to select the features most relevant to Alzheimer's disease as the radiomics features of the hippocampal volume of the target subject. Screening methods include stability screening (ICC > 0.8), maximum relevance minimum redundancy feature selection (mRMR) dimensionality reduction, and biological significance verification. It should be noted that the above is only one possible combined screening algorithm in this application, and the specific screening order and algorithm are not limited thereto.

[0072] Figure 7 The diagram shows a flowchart illustrating the training method of the Alzheimer's disease prediction model described in an embodiment of this application. Figure 7 As shown, the training method for the Alzheimer's disease prediction model includes steps S31 to S34.

[0073] Step S31: Obtain multimodal data of Alzheimer's patients, including clinical data, imaging data and radiomics data.

[0074] Step S32: Classify and label the multimodal data of Alzheimer's patients to obtain multimodal data with different risk levels.

[0075] Step S33: Preprocess the multimodal data with different risk levels to obtain the training dataset.

[0076] Step S34: Train the classification network model combined with the multi-head attention mechanism using the training dataset to obtain an Alzheimer's disease prediction model.

[0077] In some possible implementations, multimodal data of Alzheimer's patients is acquired, including clinical, imaging, and radiomics data. This multimodal data is then graded and labeled to obtain three risk levels: high-risk, intermediate-risk, and low-risk. The multimodal data at different risk levels is preprocessed to transform it into a training dataset suitable for model input. Specifically, the Least Absolute Shrinkage and Selection Operator (LASSO) and the Forward Feature Selection (FFS) algorithm are used to filter the preprocessed multimodal data to obtain the optimal combination of multimodal data as the training dataset. LASSO is a method used in statistics and machine learning for regression analysis and feature selection. The LASSO algorithm is used to filter features with non-zero coefficients, and parameters are determined through 10-fold cross-validation. The value of is determined. An XGBoost classification network model combined with a multi-head attention mechanism is used to construct a multimodal fusion prediction model for training, thereby obtaining an Alzheimer's disease prediction model. Furthermore, the trained model is validated using a multi-center external dataset, and the Alzheimer's disease prediction model is determined based on the AUC results. The Area Under the Receiver Operating Characteristic Curve (AUC) is used to evaluate the model's classification performance. External datasets include, for example, the ADNI dataset for neuroimaging, biomarkers, and early diagnosis of Alzheimer's disease, the BioFINDER dataset for early diagnosis and biomarkers of Alzheimer's disease and other neurodegenerative diseases, or other relevant datasets; this application is not limited to these.

[0078] Figure 8 The diagram shown is a structural schematic of the Alzheimer's disease prediction system described in an embodiment of this application. Figure 8 As shown, the Alzheimer's disease prediction system 100 includes a data acquisition module 110, a data processing module 120, a model prediction module 130, and a result display module 140.

[0079] The data acquisition module 110 is used to acquire multimodal data of the object under test.

[0080] The data processing module 120 is used to preprocess multimodal data and automatically quantize and calculate the perivascular space analysis index based on diffusion tensor imaging to obtain the processed test data.

[0081] The model prediction module 130 is used to predict the processed test data using an Alzheimer's disease prediction model in order to obtain the Alzheimer's disease risk level of the test subject.

[0082] The results display module 140 is used to generate visual decision reports for the test objects based on risk levels.

[0083] It should be noted that the Alzheimer's disease prediction system 100 includes the aforementioned modules 110 to 140 and Figure 3 Steps S11 to S14 in the Alzheimer's disease prediction method shown correspond one-to-one, and will not be elaborated here.

[0084] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0085] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0086] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0087] This application also provides an electronic device. Figure 9The diagram shown is a structural schematic of the electronic device described in an embodiment of this application. Figure 9 As shown, in this embodiment, the electronic device 900 includes a memory 910 and a processor 920.

[0088] The memory 910 is used to store computer programs; preferably, the memory 910 includes various media that can store program code, such as ROM, RAM, magnetic disk, USB flash drive, memory card or optical disk.

[0089] Specifically, memory 910 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. Electronic device 900 may further include other removable / non-removable, volatile / non-volatile computer system storage media. Memory 910 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this application. It is understood that memory 910 may be volatile memory or non-volatile memory, or both. Non-volatile memory may be read-only memory (ROM) or programmable read-only memory (PROM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM) and synchronous static random access memory (SSRAM). The memories described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable categories of memories.

[0090] The processor 920 is connected to the memory 910 and is used to execute the computer program stored in the memory 910 so that the electronic device 900 performs the Alzheimer's disease prediction method described in any embodiment of this application.

[0091] Optionally, the processor 920 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0092] Optionally, the electronic device 900 in this embodiment may further include a display 930. The display 930 is communicatively connected to the memory 910 and the processor 920, and is used to display the relevant graphical user interface (GUI) of the Alzheimer's disease prediction method described in this application embodiment.

[0093] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the program implements the Alzheimer's disease prediction method described in any embodiment of this application. Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0094] As used in this specification, the terms "component," "module," "system," etc., are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, and / or a computer. As illustrated, applications running on computing devices and computing devices can both be components. One or more components may reside in a process and / or an execution thread, and components may be located on a single computer and / or distributed among two or more computers. Furthermore, these components can be executed from various computer-readable media on which various data structures are stored. Components can communicate, for example, via local and / or remote processes based on signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system, and / or a network, such as the Internet interacting with other systems via signals).

[0095] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0096] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for predicting Alzheimer's disease, characterized by, The method comprises the following steps: acquiring multi-modal data of a to-be-tested object; preprocessing the multi-modal data and automatically quantifying a diffusion tensor imaging-based perivascular space analysis index to obtain processed to-be-tested data; using an Alzheimer's disease prediction model to predict the processed to-be-tested data to obtain an Alzheimer's disease risk level of the to-be-tested object; generating a visual decision report for the to-be-tested object based on the Alzheimer's disease risk level.

2. The Alzheimer's disease predicting method according to claim 1, characterized in that, The automatic quantification of the diffusion tensor imaging-based perivascular space analysis index comprises the following steps: preprocessing the diffusion tensor imaging by using a principal component analysis method to obtain corrected diffusion tensor imaging; estimating a diffusion tensor model based on the corrected diffusion tensor imaging; spatially registering the diffusion tensor model by using a standard space to obtain initial coordinates of a region of interest; calculating the distribution of the projection component and the joint component of the initial coordinates according to the tensor to obtain a peak position of the region of interest; calculating based on the peak position to obtain the diffusion tensor imaging-based perivascular space analysis index.

3. The Alzheimer's disease predicting method according to claim 2, characterized in that, The calculation based on the peak position to obtain the diffusion tensor imaging-based perivascular space analysis index comprises the following steps: generating a region of interest with a preset radius with the peak position as the center; calculating based on the region of interest to obtain the mean value of the tensor in three directions as the diffusion tensor imaging-based perivascular space analysis index.

4. The Alzheimer's disease predicting method according to claim 1, characterized in that, The multi-modal data of the to-be-tested object comprises clinical data, imaging data and imageomics data; The clinical data comprises the age, gender, education level, cognitive score, apolipoprotein E genotype and plasma p-tau217 concentration of the to-be-tested object; The imaging data comprises the hippocampal volume and the diffusion tensor imaging-based perivascular space analysis index of the to-be-tested object; The imageomics data comprises the imageomics features of the hippocampal volume of the to-be-tested object.

5. The Alzheimer's disease predicting method according to claim 4, characterized in that, The acquisition of the imageomics features of the hippocampal volume of the to-be-tested object comprises the following steps: acquiring the hippocampal volume of the to-be-tested object by using a three-dimensional-T1 weighted magnetization preparation fast gradient echo sequence; segmenting the hippocampal volume by using a hybrid segmentation algorithm to obtain hippocampal subregions; performing intracranial volume normalization processing on the hippocampal subregions to obtain residual volumes; correcting the hippocampal volume of the to-be-tested object by using the residual volumes to obtain a corrected hippocampal volume; extracting features from the corrected hippocampal volume to obtain the imageomics features of the hippocampal volume of the to-be-tested object.

6. The Alzheimer's disease predicting method according to claim 5, characterized in that, The acquisition of the imageomics features of the hippocampal volume of the to-be-tested object further comprises the following steps: combining stability screening, maximum correlation minimum redundancy feature selection algorithm dimension reduction and biological significance verification to perform multiple screening on the extracted hippocampal volume image features to obtain the imageomics features of the hippocampal volume of the to-be-tested object.

7. The Alzheimer's disease predicting method according to claim 1, characterized in that, The training method of the Alzheimer's disease prediction model comprises the following steps: acquiring multi-modal data of Alzheimer's disease patients, wherein the feature data comprises clinical data, imaging data and imageomics data; The multi-modal data of the Alzheimer's disease patient is hierarchically labeled to obtain multi-modal data with different risk levels; The multi-modal data with different risk levels is preprocessed to obtain a training data set; The classification network model combined with the multi-head attention mechanism is trained using the training data set to obtain the Alzheimer's disease prediction model.

8. An Alzheimer's disease prediction system characterized by, It comprises: a data acquisition module for acquiring multi-modal data of a to-be-tested object; a data processing module for preprocessing the multi-modal data and automatically quantifying and calculating the perivascular space analysis index based on diffusion tensor imaging to obtain processed to-be-tested data; a model prediction module for predicting the processed to-be-tested data using the Alzheimer's disease prediction model to obtain the risk level of the to-be-tested object for Alzheimer's disease; a result display module for generating a visual decision report for the to-be-tested object based on the risk level.

9. An electronic device, comprising: The electronic device comprises: a memory for storing a computer program; a processor for executing the computer program stored in the memory to enable the electronic device to perform the Alzheimer's disease prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the Alzheimer's disease prediction method according to any one of claims 1 to 7.

Images