System for patient stratification and prognosis risk stratification in patients with takayasu arteritis

Abstract

The present disclosure relates to a system for classifying and stratifying the prognosis risk of a patient with Takayasu's arteritis. The system comprises: an input module for inputting clinical data of a target patient with Takayasu's arteritis into a comprehensive scoring model; wherein the comprehensive scoring model is based on multi-modal baseline data of a plurality of untreated, initial treatment Takayasu's arteritis patients; a first determination module for determining the disease classification of the target patient with Takayasu's arteritis based on the comprehensive scoring model input by the input module; and a second determination module for determining the prognosis risk of the target patient with Takayasu's arteritis based on the disease classification determined by the first determination module. The present disclosure solves the problem that the prior art lacks quantifiable and clinically popular classification and prognosis stratification tools in the diagnosis and treatment of Takayasu's arteritis, resulting in a lack of personalized guidance for treatment options.

Description

Technical Field

[0001] This disclosure relates to the field of smart healthcare technology, and in particular to a system for classifying and stratifying the prognostic risk of patients with aortitis. Background Technology

[0002] Takayasu arteritis (TAK), a chronic autoimmune vasculitis, is seeing steady progress in clinical diagnosis and research towards precision, targeting, and systematization. The core trends are concentrated in four dimensions: innovation in diagnostic technology, optimization of treatment plans, discovery of biomarkers, and construction of personalized management systems.

[0003] Currently, the clinical treatment of Takayasu arteritis and existing related technologies still have significant limitations: traditional disease classification often relies on single inflammatory markers (such as ESR (erythrocyte sedimentation rate) or CRP (C-reactive protein)) or subjective symptom assessments by doctors, which cannot fully reflect the inherent pathophysiological heterogeneity of the disease, leading to ambiguous subtype classifications; prognostic judgment lacks precise quantitative evidence, making it impossible to effectively identify patients with high relapse, refractory, or high mortality risks in the early stages of treatment, which may delay treatment for high-risk patients due to untimely intervention, while low-risk patients may suffer unnecessary side effects from immunosuppressive drugs; at the same time, treatment plans are often "one-size-fits-all" and lack personalized guidance based on individual patient pathological characteristics, and existing technologies have not formed simple, easy-to-operate, and widely applicable clinical tools, making it difficult to support primary healthcare institutions in carrying out standardized classification and prognostic assessments, and the overall accuracy and efficiency of diagnosis and treatment need to be improved.

[0004] In summary, existing technologies lack quantifiable and clinically applicable classification and prognostic stratification tools for the diagnosis and treatment of Takayasu arteritis, resulting in a lack of personalized guidance for treatment plans. Summary of the Invention

[0005] To address, or at least partially address, the aforementioned technical problems, this disclosure provides a system for classifying and stratifying the prognostic risk of patients with aortitis.

[0006] This disclosure provides a system for classifying and stratifying the prognostic risk of patients with Takayasu arteritis. The system includes an input module for inputting clinical data of the target Takayasu arteritis patient into a comprehensive scoring model. The comprehensive scoring model is obtained based on multimodal baseline data from multiple untreated, newly diagnosed Takayasu arteritis patients. The first determining module is used to determine the disease type of the target Takayasu arteritis patient based on the comprehensive scoring model input by the input module; The second determining module is used to determine the prognostic risk of the target aortitis patient based on the disease classification determined by the first determining module.

[0007] Optionally, the system further includes: The acquisition module is used to acquire multimodal baseline data from multiple untreated, newly diagnosed patients with Takayasu arteritis; wherein the multimodal baseline data includes at least: serum cytokines detected simultaneously by flow cytometry, as well as clinical and laboratory indicators of disease activity scores, systemic inflammatory markers, immunoglobulins, complement levels, and vascular imaging data.

[0008] Optionally, the serum cytokines include at least IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17, TNF-α, IFN-α, and IFN-γ.

[0009] Optionally, the system further includes: The third determination module is used to determine at least three disease subtypes with different immune inflammatory characteristics based on the serum cytokines collected by the acquisition module and the clinical variables of the newly diagnosed aortitis patients, using an unsupervised clustering analysis algorithm. The analysis module is used to perform correlation analysis between the disease subtype determined by the third determination module and the clinical prognostic data of the newly diagnosed Takayasu arteritis patients, and to determine the risk ratio and differential indicators of each disease subtype and clinical outcome. A construction module is used to construct the comprehensive scoring model based on the risk ratio determined by the analysis module and the identification index.

[0010] Optionally, the building module further includes: The cutoff values ​​and weights of each of the discrimination indicators were determined by receiver operating characteristic (ROC) curve analysis. The comprehensive scoring model is constructed based on the cutoff value and the weight.

[0011] Optionally, the first determining module further includes: Based on the comprehensive scoring model, the disease score of the target Takayasu arteritis patient is determined; Based on the disease score and classification rules, the disease type of the target Takayasu arteritis patient is determined.

[0012] Optionally, the first determining module further includes: In response to the disease score being less than a first threshold, the disease type of the target Takayasu arteritis patient is determined to be low-risk. In response to the disease score being greater than or equal to the first threshold and less than the second threshold, the disease type of the target Takayasu arteritis patient is determined to be acute inflammatory type; wherein the second threshold is greater than the first threshold; In response to the disease score being greater than or equal to the second threshold, the disease subtype of the target Takayasu arteritis patient is determined to be multi-cytokine activated type.

[0013] Optionally, the system further includes: The validation module is used to validate the comprehensive scoring model and classification rules using a random forest classifier and a decision tree model.

[0014] Optionally, the verification module further includes: The dataset of newly diagnosed Takayasu arteritis patients was divided into a training set and a test set; The random forest classifier is trained using the training set. The classification accuracy of the comprehensive scoring model was verified on the test set using the trained random forest classifier and the decision tree model.

[0015] Optionally, the system further includes: The update module is used to iteratively optimize the parameters of the comprehensive scoring model or the classification rules based on the newly added follow-up prognostic data of newly diagnosed patients with Takayasu arteritis.

[0016] Compared with the prior art, the technical solution provided in this disclosure has the following advantages: The system imports clinical data of the target patient through the input module and connects to a comprehensive scoring model constructed based on multimodal baseline data of newly treated patients. After the first determination module accurately determines the disease type, the second determination module outputs the prognostic risk. The core beneficial effect is that it breaks through the limitations of traditional reliance on a single biomarker or subjective assessment. The comprehensive scoring model based on multimodal data ensures the objectivity and quantification of the classification and prognostic assessment, realizes the accurate identification of the patient's disease type and risks such as recurrence, refractory treatment, and death in the early stage of treatment, provides clear decision-making basis for clinical practice, effectively avoids the problem of over-medicalization or insufficient intervention caused by "one-size-fits-all" treatment, helps to formulate personalized treatment plans and optimize follow-up management, and ultimately improves the accuracy of diagnosis and treatment of aortitis and the prognosis of patients. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0018] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the system for classifying and stratifying the prognostic risk of patients with aortitis, as described in an embodiment of this disclosure. Detailed Implementation

[0020] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0021] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.

[0022] Figure 1 This is a schematic diagram of the system for classifying and stratifying the prognostic risk of patients with aortitis as described in an embodiment of this disclosure. Figure 1 As shown, the system 100 for classifying and stratifying the prognostic risk of patients with aortitis includes: an input module 101, a first determination module 102, a second determination module 103, an acquisition module 104, a third determination module 105, an analysis module 106, a construction module 107, a verification module 108, and an update module 109.

[0023] The acquisition module 104, as the data foundation layer of the system 100, is responsible for acquiring and preprocessing multimodal baseline data of newly diagnosed TAK patients, constructing standardized, high-quality analysis datasets, and ensuring the authenticity and reliability of all subsequent data-driven analyses.

[0024] Specifically, the data collection subjects were limited to: strictly screening newly diagnosed TAK hospitalized patients who had not received any immunosuppressive or antirheumatic drug treatment, eliminating interference from treatment on biomarkers, and ensuring that the data reflected the original pathological state of the disease.

[0025] The multimodal data collection scope includes at least the following: Immune inflammation spectrum data: simultaneous quantitative detection of 12 key serum cytokines by flow cytometry, specifically including IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17, TNF-α, IFN-α, and IFN-γ; Clinical and laboratory indicator data: covering demographic information (age, sex, etc.), disease activity scores (NIH score, ITAS2010 score, ITAS.A score), systemic inflammatory markers (ESR, CRPC reactive protein), immunoglobulin and complement levels; Vascular imaging data: Based on ultrasound, MRA (Magnetic Resonance Angiography), and CTA (Computed Tomography Angiography), the affected vascular sites (such as the aorta and branch arteries) and the extent of involvement are standardized and recorded according to the Hata-Numano classification standard.

[0026] The data collected by the acquisition module 104 is cleaned (outliers are removed and missing values ​​are filled), formatted and standardized, and finally outputs well-structured multimodal baseline data that can be directly used for analysis.

[0027] The serum cytokines and their functions (in conjunction with the immune-inflammatory background of TAK) are as follows: IL-1β (interleukin-1β): mediates acute inflammatory response and participates in the initiation of inflammation in the TAK vascular wall; IL-2 (interleukin-2): regulates T lymphocyte proliferation and activation, and is associated with immune dysregulation in TAK; IL-4 (interleukin-4): participates in Th2 immune response and affects the direction of immune cell differentiation; IL-5 (interleukin-5): acts on eosinophils and participates in the recruitment of inflammatory cells; IL-6 (interleukin-6): mediates systemic inflammatory response and is closely related to TAK disease activity; IL-8 (interleukin-8): belongs to the chemokine family and recruits… Neutrophils to diseased blood vessels; IL-10 (interleukin-10): has anti-inflammatory effects and regulates the intensity of immune responses; IL-12p70 (interleukin-12p70): induces Th1 immune responses and promotes the activation of inflammatory cells; IL-17 (interleukin-17): mediates chronic inflammation and tissue damage, and is associated with the progression of TAK vascular inflammation; TNF-α (tumor necrosis factor-α): a core pro-inflammatory factor, involved in TAK vascular wall inflammation and damage; IFN-α (interferon-α): a type I interferon, involved in immune regulation and TAK autoimmune disorders; IFN-γ (interferon-γ): a type II interferon, mediates Th1 immune responses and enhances the function of inflammatory cells.

[0028] The third identification module 105, as the subtype discovery layer of system 100, uses an unsupervised machine learning algorithm to objectively identify the inherent disease subtype of TAK patients from the standardized data output by the acquisition module 104, breaking the limitations of traditional single-indicator classification.

[0029] Specifically, based on the core feature variables selected from the multi-dimensional baseline dataset of acquisition module 104, including key cytokine levels (IL-6, IFN-γ, IL-1β, IL-2, etc.) and key clinical variables (age, disease duration, ESR, CRP), a two-stage clustering analysis unsupervised machine learning algorithm is used. Without pre-setting clinical labels, the data is grouped entirely based on its own characteristics, automatically dividing the data into three mutually exclusive disease subgroups, i.e., disease types, and clearly defining the biological characteristics of each subgroup. The disease types can be: acute inflammatory type: characterized by significantly elevated CRP, ESR, IL-6, and IFN-γ; multi-cytokine activation type: characterized by widespread elevation of multiple cytokines such as IL-1β, IL-2, IL-4, IL-10, TNF-α, and IFN-α (high-risk phenotype); and low-inflammatory type: without significant elevation of inflammatory markers (low-risk phenotype).

[0030] The third identification module 105 reveals the intrinsic heterogeneity of TAK disease and defines a previously unrecognized "multiple cytokine activation" high-risk subtype, providing a clear subtyping label for subsequent prognostic association verification and is a core link in achieving risk stratification.

[0031] Analysis module 106, as the clinical value verification layer of system 100, verifies the correlation between disease subtypes identified by the third determination module 105 and clinical outcomes, and builds a quantitative bridge between subtype characteristics and prognostic risks.

[0032] Specifically, the analysis is performed based on the disease classification output by the third determination module 105 and the clinical variables (including clinical feature data and prognostic data) that were not clustered by the data collection module 104; Clinical feature association analysis: statistically correlated the results of disease classification with vascular involvement patterns, comorbidities (hypertension, hyperlipidemia), and clinical symptoms; Prognostic association analysis: correlated the classification results with post-treatment remission rate, relapse rate, incidence of refractory diseases, incidence of adverse events, and all-cause mortality; Statistical methods: survival analysis (Kaplan-Meier) was used. The r-curve, Log-rank test, and multivariate Cox proportional hazards regression model were used to quantify the hazard ratio (HR) between each subtype and clinical outcome, and finally output the subtype-clinical characteristic association conclusion; subtype-prognostic association conclusion (e.g., cluster 2 is a high-risk subtype with high recurrence, high refractory and high mortality, and cluster 3 is a low-risk subtype) and key identification indicators: indicators with strong prognostic association (ESR, CRP, IL-6, IL-1β, IL-2, IL-4, IL-10, IL-12p70, TNF-α, IFN-α).

[0033] Analysis module 106 confirms the clinical practical value of data-driven typing, avoiding subtype classification from remaining only at the statistical level. At the same time, it provides high-value identification indicators and quantitative risk basis for the construction module 107, ensuring the predictive accuracy of subsequent scoring models.

[0034] Module 107, serving as the clinical tool translation layer of System 100, transforms the research results (key identification indicators, risk ratios) of Module 106 into a comprehensive scoring model that can be directly operated in clinical practice, thus realizing the implementation of research findings into clinical tools.

[0035] Specifically, based on the key identification indicators output by analysis module 106 and the corresponding prognostic hazard ratios (HRs), the indicator cutoff values ​​are determined: Receiver operating characteristic (ROC) curve analysis is used to determine the optimal cutoff value (the critical value for distinguishing different subtypes) for each key identification indicator; indicator weight assignment: based on the HR value (i.e., contribution to prognosis) of each indicator, continuous variables are transformed into categorical variables and assigned corresponding weights (e.g., 0, 1.5, 2 points), with the indicator contributing the most to identifying high-risk subtypes assigned 2 points; a comprehensive scoring model is constructed: integrating all key identification indicators, cutoff values, and weights, a comprehensive scoring model is created, with the highest disease score being 16 points; The disease type of the target patient is determined by setting thresholds. For example, classification thresholds are set according to the disease score. Assuming the first threshold is 4.75 and the second threshold is 7.25, if the disease score is < 4.75, the target patient is determined to be low-risk; if 4.75 ≤ disease score < 7.25, the target patient is determined to be acute inflammatory; if the disease score is ≥ 7.25, the target patient is determined to be high-risk multi-cytokine activation type.

[0036] Module 107 simplifies complex multi-factor clustering results into a clinical tool with linear scoring and a clear classification threshold. It can be quickly applied without professional algorithm knowledge, realizing the transformation from scientific research discovery to clinical tools and lowering the threshold for clinical promotion.

[0037] The verification module 108 serves as the reliability assurance layer of the system 100. It uses a random forest classifier and a decision tree model to verify the comprehensive scoring model and classification rules output by the construction module 107, ensuring the robustness and accuracy of the model.

[0038] Specifically, the validation process was conducted based on the comprehensive scoring model and classification threshold rules output by the construction module 107, and the standardized multimodal baseline dataset output by the acquisition module 104. The validation process included: Data partitioning: The standardized dataset of newly diagnosed aortitis patients was divided into training and test sets in a 70:30 ratio; Model training: The random forest classifier was trained using the training set; Accuracy validation: The classification accuracy of the comprehensive scoring model was validated on the test set using the trained random forest classifier and decision tree model, ultimately achieving a classification accuracy of 85.2% and a multi-class AUC of 0.964; Key feature identification: Through variable importance analysis, ESR, total score, and IL-2 were identified as core features for distinguishing disease subtypes; Rule extraction: Intuitive "if-then" classification rules were generated based on the key features, such as "If the ESR score is ≥1.5, it is classified as acute inflammatory type; if the IL-2 score is >0.75, it is classified as multi-cytokine activation type."

[0039] The validation module 108 avoids overfitting of the comprehensive scoring model, enhances the model's adaptability and reliability in new clinical samples, and provides a clear basis for the simplified classification rules output by the model to be automatically embedded into the clinical system.

[0040] The update module 109, as the dynamic optimization layer of system 100, iterates and optimizes the parameters or classification rules of the comprehensive scoring model based on the newly added follow-up prognostic data of newly diagnosed patients with aortitis, to ensure that the system adapts to actual clinical needs in the long term.

[0041] Specifically, based on the newly added multimodal baseline data, treatment follow-up data, and long-term follow-up prognostic data (including clinical outcome information such as relapse, refractory, and death) of newly diagnosed patients with Takayasu arteritis, the new data is integrated into the original dataset. The hazard ratios of each key identification indicator are recalculated, the optimal cutoff values ​​and weights of the indicators are adjusted, or the classification threshold rules are optimized. The optimized comprehensive scoring model needs to be validated again by the validation module 108 before replacing the original model.

[0042] The updated module 109 addresses the potential limitations of the initial model's limited sample size, enabling the comprehensive scoring model to continuously improve in accuracy and adaptability as clinical data accumulates, thereby extending the system's lifespan and clinical application value.

[0043] Input module 101, as the clinical application entry layer of system 100, provides a standardized data input channel for target patients with aortitis, enabling the connection between clinical test data and the comprehensive scoring model.

[0044] Specifically, target patients with Takayasu arteritis (preferably treatment-naïve patients who have not received immunosuppression or antirheumatic drug therapy) will be identified based on their type and prognostic risk stratification. Clinical data from these target Takayasu arteritis patients will be input, specifically the results of key differential diagnostic indicators (ESR, CRP, IL-6, IL-1β, IL-2, IL-4, IL-10, IL-12p70, TNF-α, IFN-α) corresponding to the comprehensive scoring model. The input clinical data will be standardized to ensure it meets the calculation requirements of the comprehensive scoring model and avoid calculation errors caused by data format differences.

[0045] Input module 101 bridges the gap between clinical test data and the comprehensive scoring model. It is the key entry point for the system to move from model building to clinical application. It is easy to operate and promote.

[0046] The first determination module 102, as the clinical classification execution layer of the system 100, determines the disease classification of the target patient based on the comprehensive scoring model and the clinical data of the target aortitis patient.

[0047] Specifically, based on the standardized clinical data of the target patient output by input module 101 and the comprehensive scoring model (including indicator weights and optimal cutoff values) output by construction module 107, the disease score is calculated: according to the weights of each key identification indicator in the comprehensive scoring model, the various test results of the target patient are scored and summed to obtain the target patient's disease score; disease classification determination: the classification is determined based on the matching relationship between the disease score and the classification threshold: in response to a disease score < the first threshold (4.75 points), the disease classification is determined to be low-inflammatory type (low-risk type); in response to a disease score ≥ the first threshold and < the second threshold (4.75 points ≤ total score < 7.25 points), the disease classification is determined to be acute inflammatory type; in response to a disease score ≥ the second threshold (7.25 points), the disease classification is determined to be multi-cytokine activation type (high-risk type).

[0048] The first determination module 102 achieves objective and standardized classification of the target patient, avoiding the errors of traditional classification that rely on doctors' subjective judgment, and provides direct classification basis for the second determination module 103.

[0049] The second determination module 103, as the clinical value output layer of the system 100, determines the prognostic risk of the target aortitis patient based on the disease classification results of the first determination module 102.

[0050] Specifically, based on the disease classification results output by the first determination module 102 and the "classification-prognostic correlation conclusion" verified by the analysis module 106, if the disease classification is low-inflammatory (low-risk), the prognostic risk is determined to be low, corresponding to low recurrence rate, low refractory rate, and low mortality rate; if the disease classification is acute-inflammatory, the prognostic risk is determined to be medium, corresponding to moderate recurrence rate, good response to conventional treatment, and low mortality rate; if the disease classification is multi-cytokine activation (high-risk), the prognostic risk is determined to be high, corresponding to high recurrence rate, high refractory rate, and high mortality rate.

[0051] The second determination module 103 transforms disease subtyping labels into clinically applicable prognostic risk assessment results, completing the closed loop of the system from data input to clinical decision support, and providing core references for doctors to formulate personalized treatment plans and follow-up strategies.

[0052] In this embodiment, for treatment-naïve TAK patients who have not received immunosuppression or antirheumatic drug treatment, by integrating multidimensional immune inflammatory markers and clinical data, and by using unsupervised learning, statistical analysis and machine learning validation techniques, objective classification of disease subtypes and precise stratification of prognostic risks are achieved. The output is quantitative results that can directly guide clinical treatment and follow-up management, and ultimately assists doctors in developing personalized treatment plans, thereby improving the accuracy and efficiency of TAK diagnosis and treatment.

[0053] The nine modules of System 100 do not operate independently; the specific logic is as follows: Model building phase: Acquisition module 104 outputs a standardized multimodal baseline dataset; Third determination module 105 outputs disease classification based on core feature variables through unsupervised clustering analysis; Analysis module 106 conducts association analysis between classification and clinical prognosis, and outputs hazard ratios and key identification indicators; Construction module 107 constructs a comprehensive scoring model and classification threshold based on the above results; Validation module 108 validates the accuracy and reliability of the model through random forest classifier and decision tree model.

[0054] Clinical application stage: Input module 101 inputs and standardizes the clinical data of the target patient; First determination module 102 calls the comprehensive scoring model to calculate the disease score and determine the disease type; Second determination module 103 outputs the prognostic risk assessment results based on the type.

[0055] Dynamic optimization phase: The update module 109 iteratively optimizes the parameters or classification rules of the comprehensive scoring model based on the newly added follow-up prognostic data of newly diagnosed Takayasu arteritis patients. After being verified by the verification module 108, the data is fed back to the clinical application phase to achieve continuous optimization of the system.

[0056] The technical solution provided in this application has the following advantages compared with the prior art: High objectivity: Through multimodal data integration and unsupervised clustering analysis, objective classification of TAK patients is achieved, avoiding interference from doctors' subjective judgment; High accuracy: The comprehensive scoring model has high AUC across multiple classes, enabling the identification of high-risk subtypes (multiple cytokine activation type), solving the pain point of existing technologies being unable to identify high-risk patients in the early stages; High practicality: Complex scientific research algorithms are transformed into a clinical tool of "linear scoring + clear thresholds," with indicators being routine testing items, making it easy to operate and quickly applicable to primary healthcare institutions; It can also be embedded in a Clinical Decision Support System (CDSS) to achieve automated classification and risk assessment, improving diagnostic and treatment efficiency; Significant value in personalized guidance: Based on the classification results, targeted treatment and follow-up suggestions are provided. High-risk patients can receive intensive treatment to reduce the risk of recurrence and death, while low-risk patients can avoid overtreatment and reduce drug side effects, achieving "one person, one policy" diagnosis and treatment; High scalability: The model is dynamically optimized through module updates, continuously improving accuracy with the accumulation of clinical data, adaptable to the diagnostic and treatment needs of different regions and populations, and has broad prospects for promotion and application.

[0057] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0058] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A system for classifying and stratifying the prognostic risk of patients with aortitis, characterized in that, The system includes: An input module is used to input clinical data of the target Takayasu arteritis patient into a comprehensive scoring model; wherein, the comprehensive scoring model is obtained based on multimodal baseline data of multiple untreated, newly diagnosed Takayasu arteritis patients; The first determining module is used to determine the disease type of the target Takayasu arteritis patient based on the comprehensive scoring model input by the input module; The second determining module is used to determine the prognostic risk of the target aortitis patient based on the disease classification determined by the first determining module.

2. The system according to claim 1, characterized in that, The system also includes: The acquisition module is used to acquire multimodal baseline data from multiple untreated, newly diagnosed patients with Takayasu arteritis; wherein the multimodal baseline data includes at least: serum cytokines detected simultaneously by flow cytometry, as well as clinical and laboratory indicators of disease activity scores, systemic inflammatory markers, immunoglobulins, complement levels, and vascular imaging data.

3. The system according to claim 2, characterized in that, The serum cytokines include at least IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17, TNF-α, IFN-α, and IFN-γ.

4. The system according to claim 1, characterized in that, The system also includes: The third determination module is used to determine at least three disease subtypes with different immune inflammatory characteristics based on the serum cytokines collected by the acquisition module and the clinical variables of the newly diagnosed aortitis patients, using an unsupervised clustering analysis algorithm. The analysis module is used to perform correlation analysis between the disease subtype determined by the third determination module and the clinical prognostic data of the newly diagnosed Takayasu arteritis patients, and to determine the risk ratio and differential indicators of each disease subtype and clinical outcome. A construction module is used to construct the comprehensive scoring model based on the risk ratio determined by the analysis module and the identification index.

5. The system according to claim 4, characterized in that, The building module also includes: The cutoff values ​​and weights of each of the discrimination indicators were determined by receiver operating characteristic (ROC) curve analysis. The comprehensive scoring model is constructed based on the cutoff value and the weight.

6. The system according to claim 1, characterized in that, The first determining module further includes: Based on the comprehensive scoring model, the disease score of the target Takayasu arteritis patient is determined; Based on the disease score and classification rules, the disease type of the target Takayasu arteritis patient is determined.

7. The system according to claim 6, characterized in that, The first determining module further includes: In response to the disease score being less than a first threshold, the disease type of the target Takayasu arteritis patient is determined to be low-risk. In response to the disease score being greater than or equal to the first threshold and less than the second threshold, the disease type of the target Takayasu arteritis patient is determined to be acute inflammatory type; wherein the second threshold is greater than the first threshold; In response to the disease score being greater than or equal to the second threshold, the disease subtype of the target Takayasu arteritis patient is determined to be multi-cytokine activated type.

8. The system according to claim 1, characterized in that, The system also includes: The validation module is used to validate the comprehensive scoring model and classification rules using a random forest classifier and a decision tree model.

9. The system according to claim 8, characterized in that, The verification module also includes: The dataset of newly diagnosed Takayasu arteritis patients was divided into a training set and a test set; The random forest classifier is trained using the training set. The classification accuracy of the comprehensive scoring model was verified on the test set using the trained random forest classifier and the decision tree model.

10. The system according to claim 1, characterized in that, The system also includes: The update module is used to iteratively optimize the parameters or classification rules of the comprehensive scoring model based on newly added follow-up prognostic data of newly diagnosed patients with Takayasu arteritis.

Images