Method, device, electronic equipment and medical system for assessing risk of neurological disease

Abstract

The present application relates to the technical field of biological information, and discloses a neural system disease risk assessment method and device, an electronic device and a medical system, the method comprising: acquiring values of a plurality of biomarkers detected based on a target gingival crevicular fluid sample; inputting the values of the plurality of biomarkers as a plurality of input features into an artificial intelligence model to obtain a probability value of a neural system disease risk. The present application comprehensively analyzes a plurality of modal biological characteristics (i.e. values of a plurality of biomarkers such as neural related proteins, pathogenic bacteria load / toxicity, inflammation / barrier, etc.) in an in-vitro gingival crevicular fluid sample, realizes intelligent auxiliary assessment of a neural system disease related risk, and the assessment result can be used for early risk screening, early warning and hierarchical management of a neural system disease, and the cost is low and there is no invasiveness.

Description

Technical Field

[0001] This invention relates to the field of bioinformatics, specifically to a method, device, electronic device, and medical system for assessing the risk of nervous system diseases. Background Technology

[0002] Neurological diseases are a class of chronic illnesses involving impairment, inflammation, or degeneration of the central nervous system, including neurodegenerative diseases such as Alzheimer's disease, encephalitis, and other neuroinflammatory diseases. These diseases often have insidious onset and a long course, and typically lack typical symptoms in the early stages. Current clinical assessments rely on imaging examinations or cerebrospinal fluid biomarker testing, but these methods have limitations such as high cost, invasiveness, and difficulty in widespread adoption for early screening. Summary of the Invention

[0003] This invention provides a method, device, electronic device, and medical system for assessing the risk of neurological diseases, in order to address the problems of high cost or invasiveness in early screening for neurological diseases.

[0004] In a first aspect, the present invention provides a method for risk assessment of neurological diseases, the method comprising:

[0005] Obtain values ​​for multiple biomarkers based on the detection of target gingival crevicular fluid samples;

[0006] The values ​​of the multiple biomarkers are used as multiple input features and input into an artificial intelligence model to obtain probability values ​​of the risk of nervous system diseases.

[0007] In a second aspect, the present invention provides a neurological disease risk assessment device, the device comprising:

[0008] The data acquisition module is used to acquire values ​​of multiple biomarkers obtained from the detection of target gingival crevicular fluid samples;

[0009] The prediction module is used to input the values ​​of the multiple biomarkers as multiple input features into an artificial intelligence model to obtain the probability value of the risk of nervous system diseases.

[0010] Thirdly, the present invention provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the neurological disease risk assessment method described in the first aspect or any corresponding embodiment thereof.

[0011] Fourthly, the present invention provides a medical system comprising an electronic device according to the third aspect above or any corresponding embodiment thereof.

[0012] Fifthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform the neurological disease risk assessment method described in the first aspect or any corresponding embodiment thereof.

[0013] In a sixth aspect, the present invention provides a computer program product, including computer instructions for causing a computer to execute the neurological disease risk assessment method described in the first aspect or any corresponding embodiment thereof.

[0014] The present invention provides a method, apparatus, electronic device, and medical system for assessing the risk of neurological diseases. By comprehensively analyzing multimodal biomarkers (such as neurorelated proteins, pathogen load / virulence, and inflammation / barrier values) in in vitro gingival crevicular fluid samples, it achieves intelligent auxiliary analysis of neurological diseases. The analysis results serve as auxiliary assessments for early risk screening, warning, and stratified management of neurological diseases, with low cost and non-invasiveness. The present invention can be widely applied in research and risk management of neurodegenerative diseases (such as Alzheimer's disease) and neuroinflammatory diseases. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention;

[0017] Figure 2 This is a flowchart illustrating a method for assessing the risk of neurological diseases according to an embodiment of the present invention;

[0018] Figure 3 This is a structural block diagram of a neurological disease risk assessment device according to an embodiment of the present invention;

[0019] Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0022] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0023] As an optional application scenario of this invention, such as Figure 1 As shown, the neurological disease risk assessment system may include at least one terminal device and at least one server. Figure 1 The system is illustrated in the example, which includes a computer 101, a mobile terminal 102, and a server 103, and the terminal devices such as the computer 101 and the mobile terminal 102 are connected to the server 103 through a network 110.

[0024] Specifically, the terminal device can be a smartphone, tablet, laptop, PDA, desktop computer, game console, smart TV, smart wearable device, in-vehicle terminal, VR (Virtual Reality) device, AR (Augmented Reality) device, etc. Server 103 can be a standalone physical server, a server cluster, a distributed system, or a cloud server providing cloud services. Network 110 can be a wired or wireless network, examples of which include, but are not limited to, the Internet, corporate intranet, local area network, wide area network, mobile communication network, and combinations thereof.

[0025] Recent studies have shown that oral microecological imbalance is closely related to central nervous system inflammation. Virulence factors (such as gingival proteases) produced by periodontal pathogens such as *Porphyromonas gingivalis* can induce neuroinflammatory responses, promote abnormal deposition of β-amyloid (Aβ), and participate in the neurological damage process through the blood-brain barrier (BBB) ​​pathway. Gingival crevicular fluid (GCF) is the exudate from the gingival sulcus, originating from periodontal capillaries and interstitial spaces. It contains host immune proteins, bacterial products, and inflammatory mediators. Existing literature has confirmed a significant correlation between inflammatory factors, metabolites, and neuroprotein markers in GCF and the progression of neurological diseases. Compared to blood or cerebrospinal fluid, GCF collection is simple, non-invasive, and repeatable, giving it a natural advantage as a body fluid biomarker for screening.

[0026] According to an embodiment of the present invention, a method for risk assessment of neurological diseases is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] This embodiment provides a method for assessing the risk of neurological diseases, which can be used in various electronic devices, such as the aforementioned terminal devices or servers. Figure 2 This is a flowchart of a neurological disease risk assessment method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps:

[0028] Step S201: Obtain the values ​​of multiple biomarkers based on the detection of the target gingival crevicular fluid sample.

[0029] Step S202: The values ​​of the multiple biomarkers are used as multiple input features and input into the artificial intelligence model to obtain the probability value of the risk of nervous system diseases.

[0030] Among them, the target gingival crevicular fluid sample is the gingival crevicular fluid sample of the target object, and the probability value of the neurological disease risk refers to the probability value of the target object being in a neurological disease-related abnormal state.

[0031] The neurological disease risk assessment method provided in this embodiment comprehensively analyzes multimodal biomarkers (such as neurorelated proteins, pathogen load / virulence, inflammation / barrier, etc.) in in vitro gingival crevicular fluid samples to achieve intelligent auxiliary analysis of neurological diseases. The analysis results serve as an auxiliary assessment for early risk screening, early warning, and stratified management of neurological diseases, and are low-cost and non-invasive. Specifically, this embodiment uses non-invasive oral sampling, eliminating the need for expensive or invasive examinations such as lumbar puncture and PET-CT. GCF collection is repeatable and low-cost, making it suitable for early risk screening or follow-up monitoring of at-risk populations.

[0032] The risk assessment method for neurological diseases provided in this embodiment can be widely applied in the research and risk management of neurodegenerative diseases, neuroinflammatory diseases, and other fields.

[0033] It should be noted that the analysis results obtained by this method are for doctors' reference only, and the final diagnosis should be based on the doctor's diagnosis.

[0034] In this embodiment, gingival crevicular fluid samples were collected using a standardized non-invasive method. Specifically, the target subject needed to fast, abstain from water, and refrain from oral cleaning for one hour prior to the sample collection. Then, through periodontal examination, 3-5 sites with a periodontal pocket depth ≥4mm were selected as sampling points. Finally, a sterile filter paper strip (e.g., a gingival crevicular fluid sampling strip, also known as a PeriodoPaper absorbent strip) was inserted into the gingival sulcus of the target subject and left for 30 seconds for absorption. Any samples contaminated with blood visible to the naked eye were excluded, thus completing the standardized non-invasive collection.

[0035] In addition, after collecting gingival crevicular fluid samples from multiple periodontal sites, the filter paper strips from all sites were combined into a single centrifuge tube, and a quantitative elution buffer was added. The gingival crevicular fluid was then eluted from the filter paper strips through vortexing and centrifugation. The total eluted gingival crevicular fluid sample was immediately aliquoted into multiple sample tubes and placed in a sterile filter paper cassette. These tubes were then picked up by a robotic arm for testing on different detection platforms to obtain values ​​(i.e., test values) for multiple biomarkers.

[0036] In this embodiment, multiple biomarkers are characteristic and related risk markers corresponding to neurological diseases.

[0037] In some optional embodiments, the plurality of biomarkers includes at least one of the following classes:

[0038] bacteria;

[0039] Bacterial metabolic products;

[0040] Inflammatory factors;

[0041] Nervous system-related protein biomarkers.

[0042] The bacteria include at least one of the following: Porphyromonas gingivalis; Fusobacterium nucleatum; Treponema denticola; T. forsythia; and Aggregates actinomycetemcomitans.

[0043] Among them, *Porphyromonas gingivalis* was quantitatively detected based on the 16S rRNA gene and the fimbrial A (fimA) virulence gene. The abundance of a total of six microbial targets could be detected simultaneously using multiplex quantitative PCR technology.

[0044] The bacterial metabolites include at least one of the following: butyrate, isobutyrate, propionate, gingipain Rgp / Kgp, and lipopolysaccharide (P. gingivalis-LPS).

[0045] The inflammatory factors include at least one of the following: interleukin-6 (i.e., IL-6), matrix metalloproteinase-9 (MMP-9), chemokine ligand 10 (CXCL10), and interleukin-1β (IL-1β).

[0046] Neurological markers include at least one of the following: β-amyloid protein (including Aβ42 and Aβ40), phosphorylated Tau protein (specifically p-tau217), glial fibrillary acidic protein (GFAP), and neurofilament light chain protein (NfL).

[0047] For example, when the neurological disease is a neurodegenerative disease, the plurality of biomarkers include:

[0048] Neurologically related proteins: Aβ-related markers (Aβ42, Aβ40, Aβ42 / 40 ratio), where an Aβ-related marker value ≥1 pg / mL in gingival crevicular fluid is considered abnormal, and an Aβ42 / 40 ratio ≤0.10 is also considered abnormal;

[0049] Pathogen load / virulence: Porphyromonas gingivalis ( Porphyromonas gingivalis, P. gingivalis ) load, gingipain content, among which P. gingivalis Load detection value ≥1×10 4 / site is abnormal; a gingipain content detection value ≥1 ng / site is abnormal.

[0050] Inflammation / barrier: MMP-9, IL-6, where MMP-9 levels ≥20 ng / mL / 30 s are considered abnormal, and IL-6 levels ≥1 ng / mL are considered abnormal.

[0051] When the neurological disease is a neuroinflammatory disease (i.e., a neuroinflammatory disease), the plurality of biomarkers include:

[0052] Inflammation / barrier: IL-6, MMP-9, CXCL10, where IL-6 ≥1 ng / mL is abnormal, MMP-9 ≥20 ng / mL / 30 s sample is abnormal, and CXCL10 ≥60 pg / mL is abnormal;

[0053] Pathogen load / virulence: P. gingivalis Load, gingipain content, among which P. gingivalis Load detection value ≥1×10 4 / site is abnormal; a gingipain content detection value ≥1 ng / site is abnormal.

[0054] Specifically, a real-time quantitative polymerase chain reaction (qPCR) platform can be used to detect bacterial load and inflammatory factors; an immunoassay platform can be used to detect the concentration of neurological-related protein markers and bacterial metabolites, specifically using high-sensitivity multiplex immunoassay techniques (such as Luminex liquid chromatography-mass spectrometry) to simultaneously quantify multiple protein markers.

[0055] After the tests are completed on each testing platform, the raw data output from different platforms can be preprocessed. Specifically, this includes: converting the Ct (Cycle Threshold) values ​​output from the qPCR testing platform into relative quantitative values ​​using the 2^(-ΔΔCt) method; performing a logarithmic transformation (Log(x+1)) on the concentration data obtained from the immunoassay platform to approximate a normal distribution; and finally, normalizing all features (the values ​​of each biomarker) using Z-scores to eliminate the influence of dimensions. Additionally, missing values ​​are imputed using the K-Nearest Neighbors (KNN) algorithm. After preprocessing, a standardized feature matrix of the input features is obtained, which can then be used to input the artificial intelligence model for neurological disease risk assessment.

[0056] The aforementioned preprocessing can be performed by the computing motherboard of a dedicated graphics processing unit (GPU) on an electronic device.

[0057] In some optional implementations, the artificial intelligence model includes: a nonlinear interaction representation module, a feature fusion module, and an output module;

[0058] The nonlinear interactive representation module is used to extract the nonlinear feature representations corresponding to each input feature;

[0059] The feature fusion module is used to perform weighted fusion of the nonlinear feature representations to obtain fused features;

[0060] The output module is used to output the probability value of the risk of the nervous system disease based on the fusion features.

[0061] Specifically, the nonlinear interactive representation module is a Lightweight Gradient Boosting Machine (LightGBM).

[0062] Furthermore, the feature fusion module can be implemented using a multilayer perceptron (MLP). This MLP can contain two fully connected hidden layers (64 and 32 neurons respectively). The output module uses a single-neuron layer with a sigmoid activation function to output the final risk probability value, which is a continuous probability value between 0 and 1. Additionally, the artificial intelligence model includes an input module responsible for loading the standardized feature matrix.

[0063] In this embodiment, nonlinear interaction features can be captured using LightGBM, and then deeply weighted and fused through an MLP layer to output a probability value (i.e., a personalized risk index) Y for the neurological disease risk of the target object, expressed by the formula:

[0064] Y = σ(Σ w i · f i (x i ) + b)

[0065] Where, x i Let f represent the i-th input feature. i (x i Let w be the nonlinear representation of the i-th input feature extracted by LightGBM. i For feature weights calculated based on mutual information, w i = I(F i ; Y) / Σ I(F j ; Y), reflecting its contribution to the prediction result Y, b is the bias term, and σ is the Sigmoid function, used to map the model output to a continuous probability value in the interval of 0 to 1.

[0066] If the probability value of the risk of neurological diseases to be predicted is the Alzheimer's disease risk index, then the predicted result Y is also the Gingival Crevicular Fluid–AD Risk Index (GARI).

[0067] In this embodiment, the artificial intelligence model used to predict the probability value of neurological disease risk adopts a hybrid architecture combining LightGBM and Multilayer Perceptron (MLP). LightGBM is responsible for capturing high-order nonlinear interaction relationships between input features of different modalities (achieved through nonlinear transformation and high-order interaction learning), mining potential synergistic or antagonistic effects, and outputting a new set of nonlinear feature representations rich in interaction information. MLP is responsible for deep feature integration and weighted modeling to further weight and fuse and abstract the deep features extracted by LightGBM, forming a unified representation of neurological diseases. The network structure is [20 → 64 → 32 → 1], the activation function is ReLU, and Dropout is set to 0.2 to prevent overfitting and enhance generalization ability.

[0068] In summary, the artificial intelligence model provided in this embodiment achieves unified quantification of multimodal input features through "feature importance weighting + nonlinear interactive representation + deep fusion", which gives greater weight to input features with higher information content, improves the model's expressive power, stability and interpretability, and avoids bias in analysis caused by features of different dimensions and types.

[0069] The above prediction process of the artificial intelligence model can be implemented by a hardware acceleration card of a field-programmable gate array (FPGA) chip.

[0070] In some optional implementations, the artificial intelligence model further includes a SHAP (SHapley Additive ex Planations) interpretability analysis module, which is used to calculate the contribution of each input feature to the probability value predicted by the artificial intelligence model and quantify its positive or negative impact.

[0071] In this embodiment, the artificial intelligence model used for assessing the risk of neurological diseases also integrates the SHAP interpretation framework. This framework enables attribution analysis of the risk probability values ​​output by the model, quantifies the contribution of each input feature to the prediction results, and outputs risk distribution maps and feature importance visualization maps (specifically, contribution bar charts). This allows for the generation of interpretable biomarker analysis reports (e.g., gingipain content, MMP-9 concentration, and Aβ-related biomarker content have the highest SHAP value contribution in high-risk samples, suggesting that the inflammation-neural axis signaling pathway plays an important role in the early formation of Alzheimer's disease (AD)). This enhances model transparency, enabling researchers and clinicians to understand how oral-neural axis signals affect the risk of neurological diseases, facilitating scientific validation and clinical understanding, and supporting clinical decision-making.

[0072] After using an artificial intelligence model to complete the risk assessment of neurological diseases and obtain the contribution of each input feature to the prediction results, a risk assessment report can be generated. This risk assessment report includes: the overall risk index (i.e., the probability value of the risk of neurological diseases), the risk level, the ranking of the contribution of key features, health tips, clinical recommendations, and other content.

[0073] The risk level can be determined based on the probability value of the risk of neurological diseases predicted by the artificial intelligence model. Specifically, the risk can be divided into three levels according to the probability value of the risk of neurological diseases: low risk (risk probability value < 0.3), medium risk (0.3 ≤ risk probability value ≤ 0.6), and high risk (risk probability value > 0.6). This threshold can be adjusted according to the sensitivity requirements of the screening scenario.

[0074] The following example illustrates the training and validation process of an artificial intelligence model for risk assessment of neurological diseases: predicting the risk of recurrence of optic neuritis within one year based on multimodal biomarkers of gingival crevicular fluid.

[0075] 1. Research Subjects and Sample Acquisition

[0076] This embodiment is based on an ethically approved clinical study that consecutively enrolled 30 patients with optic neuritis. Gingival crevicular fluid (GCF) samples were collected from the target subjects at baseline (first visit or within 7 days after acute remission). Gingival crevicular fluid collection was performed using a standardized method of placing a filter paper strip at the target site for 30 seconds; the sampling site was the mesiobuccal site of the maxillary first molar.

[0077] All patients were followed up for 12 months after sampling. Recurrence was determined based on clinical symptoms, imaging findings, and physician diagnosis. Patients who experienced recurrence within 12 months were assigned to the recurrence group (R group), and those who did not relapse were assigned to the non-recurrence group (NR group). In this example, there were 11 patients (36.6%) in the recurrence group and 19 patients (63.4%) in the non-recurrence group.

[0078] 2. Biomarker Detection and Anomaly Threshold Definition

[0079] The multimodal biomarkers of gingival crevicular fluid detected in this embodiment include:

[0080] (1) Inflammation / barrier-related biomarkers: IL-6, MMP-9, CXCL10. The detection method was Luminex liquid chromatography-mass spectrometry. The abnormal thresholds were defined as follows: IL-6: a value ≥ 1 ng / mL was considered abnormal; MMP-9: a value ≥ 20 ng / mL / 30s was considered abnormal; CXCL10: a value ≥ 60 pg / mL was considered abnormal.

[0081] (2) Pathogen load / virulence-related markers: P. gingivalis Load, gingipain content. P. gingivalis Loading was detected by qPCR; gingipain content was detected by ELISA and expressed as ng / site. Abnormal thresholds were defined as follows: P. gingivalis Load: Detected value ≥ 1×10 4 / site is considered abnormal; gingipain content: a detection value ≥ 1 ng / site is considered abnormal.

[0082] Note: The above "abnormalities" are used to characterize abnormal infection load / virulence load and inflammation-barrier damage load for risk assessment and stratification, and are not equivalent to a confirmed diagnosis.

[0083] 3. Dataset Construction and Feature Engineering

[0084] For each subject, an input feature vector is constructed, including:

[0085] (a) Continuous variables: IL-6, MMP-9, CXCL10, P. gingivalis Load, gingipain test value;

[0086] (b) Binary variables: Whether the above 5 items are abnormal (0 / 1);

[0087] (c) Optional covariates (for correction): age, gender.

[0088] This embodiment also constructs the "Gingival Crevice Fluid-ON Risk Index" (GORI): GORI = w1·I(IL-6 abnormality) + w2·I(MMP-9 abnormality) + w3·I(CXCL10 abnormality) + w4·I(Pg abnormality) + w5·I(gingipain abnormality), where the weights w1–w5 are obtained by model training or determined by information gain / feature importance.

[0089] 4. Artificial Intelligence Model Training and Validation

[0090] This embodiment uses a hybrid architecture of LightGBM + Multilayer Perceptron (MLP):

[0091] First, LightGBM is used to calculate feature importance and information gain, and key variables are selected as inputs for subsequent joint training. Then, the selected features are input into the MLP for cross-modal fusion learning, and the relapse risk probability value (0–1) is output.

[0092] Considering a sample size of 30 cases, including 11 cases in the relapse group and 19 cases in the non-relapse group, this embodiment employs stratified 5-fold cross-validation: while ensuring a relatively consistent relapse / non-relapse ratio across all folds, the sample is randomly divided into 5 subsets; each time, 4 folds are used as the training set and 1 fold as the validation set, and this process is repeated 5 times before the average performance index is calculated. To reduce fluctuations caused by random partitioning, the fluctuation range of the validation results for each fold is controlled within ±0.03.

[0093] Training optimization uses Adam (learning rate 0.001, batch size 8), with binary cross entropy as the loss function, and employs an early stopping mechanism (patience=10) to improve generalization performance.

[0094] 5. Results: Comparison of the abnormality rate between the non-relapse group (NR group) and the relapse group (R group).

[0095] Based on the above thresholds, the number of abnormal cases in the two groups of samples is summarized in Table 1 below (the values ​​in parentheses are proportions):

[0096] Table 1. Statistics on the number of outliers in the two groups of samples.

[0097]

[0098] It is evident that, compared to the non-relapse group, the relapse group showed abnormalities in inflammation / barrier-related markers (IL-6, MMP-9, CXCL10) and pathogen load / virulence. P. gingivali The proportions of s and gingipain were higher, suggesting that this multimodal feature is associated with the risk of recurrence within one year.

[0099] 6. Model prediction performance (cross-validation results)

[0100] A risk assessment model was constructed and validated based on the above multimodal gingival crevicular fluid characteristics. The results show that:

[0101] The mean area under the curve (AUC) was 0.89; sensitivity was approximately 0.85; specificity was approximately 0.82; and accuracy was approximately 0.84. Overall, the model outperformed single biomarker (single indicator) predictions; the fluctuation range of each validation result was less than ±0.03, indicating strong robustness and generalization ability.

[0102] 7. Conclusion

[0103] This embodiment compares the inflammatory / barrier markers (IL-6, MMP-9, CXCL10) and pathogen load / virulence markers in gingival crevicular fluid. P. gingivalis A method for intelligently assessing the risk of recurrence of optic neuritis within one year was established and validated using a fusion model of α and gingipain. The risk probability value output by the model enables stratified assessment of recurrence risk, and those with higher risk assessments have a higher recurrence rate during follow-up, thus providing technical support for early screening and warning of recurrence risk related to optic neuritis.

[0104] In summary, this invention proposes an intelligent assessment method for the risk of neurological diseases based on the fusion of multimodal features of gingival crevicular fluid. It is the first to achieve integrated analysis and intelligent risk quantification of multidimensional biomarkers of the periodontal-neural axis, and also the first to propose intelligent analysis of neurological disease risk using gingival crevicular fluid as a single body fluid sample. This method can uniformly process and perform multimodal fusion analysis of bacteria, inflammation, and neurological-related protein indicators in in vitro gingival crevicular fluid samples, providing an in vitro, non-invasive, interpretable, and scalable technical means for the auxiliary analysis of risks related to neurological diseases (such as Alzheimer's disease and inflammatory neurological diseases).

[0105] In addition, the technical solution provided by the embodiments of the present invention also has the following advantages:

[0106] I. Utilize SHAP to perform interpretability analysis, obtain the contribution of key features, and improve the model's credibility and clinical usability.

[0107] II. The testing system consists of qPCR and immunoassay, both of which are standard platforms that can be performed in routine hospital laboratories. The technical solutions are easy to adapt and can be integrated with existing testing procedures and deployed automatically.

[0108] Third, it is highly sensitive to oral microecology and immune status, and can be dynamically monitored over time. It can be used to observe changes in risk index in high-risk groups for neurological diseases after oral treatment or behavioral intervention, and has long-term follow-up value.

[0109] This embodiment also provides a neurological disease risk assessment device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0110] This embodiment provides a device for assessing the risk of neurological diseases, such as... Figure 3 As shown, it includes:

[0111] Data acquisition module 301 is used to acquire values ​​of multiple biomarkers obtained based on the detection of target gingival crevicular fluid samples;

[0112] The prediction module 302 is used to input the values ​​of the multiple biomarkers as multiple input features into an artificial intelligence model to obtain the probability value of the risk of nervous system diseases.

[0113] In some alternative implementations, the plurality of biomarkers includes at least one of the following classes:

[0114] bacteria;

[0115] Bacterial metabolic products;

[0116] Inflammatory factors;

[0117] Nervous system-related protein biomarkers.

[0118] In some alternative embodiments, the bacteria include at least one of the following: Porphyromonas gingivalis; Fusobacterium nucleatum; Treponema denticulatum; Forsythia stolonifer; Aggregates actinomycetes;

[0119] The bacterial metabolites include at least one of the following: butyric acid, isobutyric acid, propionic acid, gingival protease, and lipopolysaccharide;

[0120] The inflammatory factors include at least one of the following: interleukin-6, matrix metalloproteinase-9, chemokine ligand 10, and interleukin-1β;

[0121] Neurological markers include at least one of the following: β-amyloid protein, phosphorylated Tau protein, glial fibrillary acidic protein, and neurofibrillary light chain protein.

[0122] In some optional implementations, the artificial intelligence model includes: a nonlinear interaction representation module, a feature fusion module, and an output module;

[0123] The nonlinear interactive representation module is used to extract the nonlinear feature representations corresponding to each input feature;

[0124] The feature fusion module is used to perform weighted fusion of the nonlinear feature representations to obtain fused features;

[0125] The output module is used to output the probability value of the risk of the nervous system disease based on the fusion features.

[0126] In some alternative implementations, the nonlinear interactive representation module is a lightweight gradient booster.

[0127] In some optional implementations, the artificial intelligence model further includes a SHAP interpretability analysis module, which is used to calculate the contribution of each of the input features to the probability value predicted by the artificial intelligence model.

[0128] The neurological disease risk assessment device provided in this embodiment of the invention can execute the neurological disease risk assessment method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the various modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0129] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0130] The following is a detailed reference. Figure 4 The diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor 401, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 402 or a program loaded from a memory 408 into a random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the electronic device. The processor 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0131] Specifically, in this embodiment, the processor 401 in the electronic device may include a dedicated graphics processor (GPU) and a hardware acceleration card with a field-programmable gate array (FPGA) chip. The dedicated graphics processor (GPU) is specifically used for the preprocessing of the raw detection data in the above method embodiment. The hardware acceleration card with the field-programmable gate array (FPGA) chip can be used for the above-mentioned process of using an artificial intelligence model for risk assessment of neurological diseases and training of the artificial intelligence model.

[0132] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0133] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a memory 408, or installed from a ROM 402. When the computer program is executed by the processor 401, it performs the functions defined in the neurological disease risk assessment method of the embodiments of the present invention.

[0134] Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0135] This invention also provides a medical system, including any of the electronic devices described in the above embodiments.

[0136] In some optional embodiments, the medical system further includes at least one of a sample collection device, a detection device, and a report output device;

[0137] The sample collection device is used to collect target gingival crevicular fluid samples;

[0138] The detection device is used to detect the values ​​of multiple biomarkers obtained from the target gingival crevicular fluid sample;

[0139] The report output device is used to output a risk assessment report for nervous system diseases.

[0140] Specifically, the sample collection device is responsible for non-invasive collection of gingival crevicular fluid, sample barcoding, and electronic information registration. This device can be an integrated sampling workstation including a sterile filter paper compartment, a robotic arm gripping unit, sample tube slots, and a barcode scanner. The testing device is used to perform qPCR and immunoassay tasks, and can be a laboratory testing platform (including the aforementioned qPCR, Luminex liquid chromatography chip, etc.). The report output device is used to output the aforementioned risk assessment report (which may include: overall risk index (i.e., the probability value of the risk of neurological diseases), risk level, ranking of key feature contributions, health tips, clinical recommendations, etc.). The report output device can be a terminal device integrating a touch screen and a report printing module.

[0141] In this embodiment, the devices and the electronic devices can be connected via wired or wireless networks to realize the entire process of in vitro analysis, including sample information acquisition, signal detection, data transmission and result output.

[0142] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium after being downloaded via a network. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the neurological disease risk assessment method shown in the above embodiments is implemented.

[0143] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0144] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for risk assessment of neurological diseases, characterized in that, The method includes: Target gingival crevicular fluid samples were collected from multiple periodontal sites of the target subject; The method involves obtaining numerical values ​​for multiple biomarkers obtained from the detection of target gingival crevicular fluid samples. These biomarkers include the following categories: bacteria, bacterial metabolites, inflammatory factors, and neurologically related protein markers. The bacteria include at least one of the following: *Porphyromonas gingivalis*, *Fusobacterium nucleatum*, *Treponema denticulatum*, *Forsythia suspensa*, and *Aggregobacter actinomycetii*. The bacterial metabolites include at least one of the following: butyrate, isobutyrate, propionic acid, gingival protease, and lipopolysaccharide. The inflammatory factors include at least one of the following: interleukin-6, matrix metalloproteinase-9, chemokine ligand 10, and interleukin-1β. The neurologically related protein markers include at least one of the following: β-amyloid protein, phosphorylated Tau protein, glial fibrillary acidic protein, and neurofibrillary light chain protein. The values ​​of the multiple biomarkers are used as multiple input features and input into an artificial intelligence model to obtain the probability value of the risk of nervous system diseases. The probability value of the risk of nervous system diseases is the risk probability value of the target object being in a nervous system disease-related abnormal state. The artificial intelligence model includes: a nonlinear interaction representation module, a feature fusion module, and an output module; The nonlinear interactive representation module is used to extract the nonlinear feature representations corresponding to each input feature, and the nonlinear interactive representation module is a lightweight gradient booster. The feature fusion module is used to perform weighted fusion on the nonlinear feature representation to obtain fused features. The feature fusion module is a multilayer perceptron. The output module is used to output the probability value of the risk of the nervous system disease based on the fusion features, and the output module adopts a single neuron layer with the Sigmoid activation function.

2. The method according to claim 1, characterized in that, The artificial intelligence model also includes a SHAP interpretability analysis module, which is used to calculate the contribution of each input feature to the probability value predicted by the artificial intelligence model.

3. A device for assessing the risk of neurological diseases, characterized in that, The device includes: The data acquisition module is used to collect target gingival crevicular fluid samples from multiple periodontal sites of the target object, and to acquire values ​​of multiple biomarkers obtained based on the detection of the target gingival crevicular fluid samples. The multiple biomarkers include the following categories: bacteria, bacterial metabolites, inflammatory factors, and neurologically related protein markers. The bacteria include at least one of the following: *Porphyromonas gingivalis*, *Fusobacterium nucleatum*, *Treponema denticulatum*, *Forsythia suspensa*, and *Aggregatibacter actinomycetes*. The bacterial metabolites include at least one of the following: butyrate, isobutyrate, propionic acid, gingival protease, and lipopolysaccharide. The inflammatory factors include at least one of the following: interleukin-6, matrix metalloproteinase-9, chemokine ligand 10, and interleukin-1β. The neurologically related protein markers include at least one of the following: β-amyloid protein, phosphorylated Tau protein, glial fibrillary acidic protein, and neurofibrillary light chain protein. The prediction module is used to input the values ​​of the multiple biomarkers as multiple input features into an artificial intelligence model to obtain the probability value of the risk of neurological diseases. The probability value of the risk of neurological diseases is the risk probability value of the target object being in a neurological disease-related abnormal state. The artificial intelligence model includes: a nonlinear interactive representation module, a feature fusion module, and an output module. The nonlinear interactive representation module is used to extract the nonlinear feature representations corresponding to each input feature, and the nonlinear interactive representation module is a lightweight gradient booster. The feature fusion module is used to perform weighted fusion on the nonlinear feature representation to obtain fused features. The feature fusion module is a multilayer perceptron. The output module is used to output the probability value of the risk of the nervous system disease based on the fusion features, and the output module adopts a single neuron layer with the Sigmoid activation function.

4. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the neurological disease risk assessment method according to claim 1 or 2.

5. A medical system, characterized in that, Includes the electronic device as described in claim 4.

6. The medical system according to claim 5, characterized in that, Also includes: At least one of a sample collection device, a detection device, and a report output device; The sample collection device is used to collect target gingival crevicular fluid samples; The detection device is used to detect the values ​​of multiple biomarkers obtained from the target gingival crevicular fluid sample; The report output device is used to output a risk assessment report for nervous system diseases.

Images