A method and system for evaluating inflammatory bowel disease
By acquiring biological sample data, including barrier function indicators, inflammatory factors, and gut microbiota indicators, and performing standardized preprocessing, a comprehensive assessment index is constructed. This addresses the challenges of existing inflammatory bowel disease (IBD) assessment methods, which suffer from high subjectivity, invasiveness, and infrequent testing. The result is a non-invasive, convenient, and accurate IBD assessment that comprehensively reflects the pathophysiological state and enables dynamic risk prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-16
- Publication Date
- 2026-07-03
Smart Images
Figure CN122337628A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of medical diagnostic auxiliary technology, specifically relating to a method and system for assessing inflammatory bowel disease. Background Technology
[0002] Inflammatory bowel disease (IBD) is a chronic, nonspecific inflammatory bowel disease whose etiology and pathophysiological mechanisms are not yet fully understood. Due to its long course, high relapse rate, and progressive aggravation, accurate assessment of its severity is of core significance for disease management, treatment adjustment, and prognosis.
[0003] Existing methods for assessing inflammatory diseases mainly include clinical symptom scoring, biomarker assessment, and invasive examination. However, clinical symptom scoring is based only on macroscopic clinical symptom scores and cannot capture subtle changes in early pathology, making it highly subjective. Biomarker assessment is prone to generating one-sided assessment results. Invasive examination is invasive, easily causes pain to patients, cannot be performed frequently to achieve dynamic monitoring, and also has a high degree of subjectivity. Summary of the Invention
[0004] To overcome the shortcomings of the prior art, the present invention proposes a method for assessing inflammatory bowel disease, the method comprising:
[0005] Obtain biological sample data of the target object; the biological sample data includes at least barrier function index data, inflammatory factor index data, and gut microbiota index data. The biological sample data are preprocessed to standardize the indicator data. Based on the preprocessed biological sample data, the barrier function score, inflammation score, and gut microbiota score, which characterize the individual state of the target object, are calculated. Based on the barrier function score, the inflammation score, and the gut microbiota score, a comprehensive assessment index is constructed to quantitatively evaluate the inflammatory bowel disease of the target subjects.
[0006] Specifically, the method for preprocessing the biological sample data includes: The relative abundance data of microorganisms in the gut microbiota index data are transformed; The gene FPKM values in the barrier function index data and the inflammatory factor index data are converted; Each of the aforementioned indicator data is standardized to make the different indicator data additive.
[0007] Preferably, the transformation of the relative abundance data of microorganisms includes a central logarithmic ratio transformation; The transformations performed on the gene FPKM values include log2 transformation, log10 transformation, or Box-Cox transformation; The standardization process includes Z-Score standardization or max-min standardization.
[0008] Preferably, the barrier function index data includes at least protective data and damaging data, and the barrier function score is obtained by subtracting the mean z-score of the damaging data from the mean z-score of the protective data; The inflammatory factor index data includes at least pro-inflammatory data and anti-inflammatory data, and the inflammation score is obtained by subtracting the mean z-score of the pro-inflammatory data from the mean z-score of the anti-inflammatory data. The gut microbiota index data includes at least beneficial bacteria data and harmful bacteria data, and the gut microbiota score is obtained by subtracting the mean z-score of the harmful bacteria data from the mean z-score of the beneficial bacteria data.
[0009] Specifically, the method for constructing the comprehensive evaluation index includes: The average of the barrier function score, the inflammation score, and the gut microbiota score is determined as the comprehensive evaluation index; Alternatively, by training an elastic network model, weight coefficients can be determined for the barrier function score, inflammation score, and gut microbiota score, and the weighted calculation result of the barrier function score, inflammation score, and gut microbiota score can be determined as the comprehensive evaluation index.
[0010] Specifically, the method for training the elastic network model includes: Obtain a training dataset containing barrier function scores, inflammation scores, gut microbiota scores, and corresponding disease activity scores, Mayo scores, SES-CD scores, and / or CDEIS scores for multiple individual samples; The elastic network model is constructed by using barrier function score, inflammation score and gut microbiota score from the training dataset as input features, and disease activity score, Mayo score, SES-CD score and / or CDEIS score as prediction targets. The elastic network model is trained using leave-one-out cross-validation to optimize the mixing parameters and regularization strength of the elastic network model in order to minimize the model prediction error until the preset convergence condition is met, thus completing the training of the elastic network model.
[0011] Furthermore, the method also includes: Obtain biological sample data of the target object at multiple different time points; A comprehensive evaluation index corresponding to each of the aforementioned time points is obtained based on the biological sample data. Trend analysis is performed based on the comprehensive assessment indices to predict the progression of inflammatory bowel disease, treatment efficacy, and / or recurrence risk in the target subjects.
[0012] Furthermore, the method also includes generating an evaluation result after constructing the comprehensive evaluation index, the evaluation result including: The barrier function score, inflammation score, microbial score, and comprehensive assessment index at a single time point, as well as the severity grading of inflammatory bowel disease corresponding to the comprehensive assessment index; And / or, display radar charts showing the barrier function score, inflammation score, microbial score, and comprehensive assessment index for a single time point; And / or, a dynamic risk prediction report generated based on a comprehensive assessment index corresponding to multiple time points, including disease trends, probability of worsening, probability of remission, and / or treatment recommendations.
[0013] Preferably, the method further includes: Divide multiple different target objects into several groups; Based on the barrier function score, inflammation score, and gut microbiota score of all target subjects in each group, the average barrier function score, average inflammation score, and average gut microbiota score of each group are calculated to achieve a comparison of population differences between different groups.
[0014] The present invention also proposes an assessment system for inflammatory bowel disease, the system comprising: The acquisition module is used to acquire biological sample data of the target object; the biological sample data includes at least barrier function index data, inflammatory factor index data, and gut microbiota index data. The processing module is used to preprocess the biological sample data to eliminate dimensional differences between the various indicator data; The calculation module is used to calculate the barrier function score, inflammation score, and gut microbiota score, which characterize the individual state of the target object, based on the preprocessed biological sample data. The assessment module is used to construct a comprehensive assessment index for quantitatively assessing the inflammatory bowel disease of the target subject based on the barrier function score, the inflammation score, and the gut microbiota score.
[0015] The present invention has at least the following beneficial effects: The proposed solution integrates and evaluates multiple indicators, and generates a quantifiable comprehensive evaluation index through standardized preprocessing, thereby comprehensively capturing the pathophysiological state of inflammatory bowel disease. The entire process does not rely on patient complaints or observer subjective judgment, effectively overcoming the problems of strong subjectivity and large differences between different assessors in existing scoring methods, avoiding misdiagnosis or missed diagnosis caused by single-dimensional assessment, and achieving accurate differentiation between potential and overt diseases. Furthermore, the proposed solution transforms the relative abundance data of microorganisms, making the variance of the relative abundance data more stable and ensuring the effectiveness of subsequent standardization processing. This solution draws on the "immune balance theory" and unifies the indicator direction of each indicator data by synergistically adjusting the calculation method of each score, so that the subsequent score calculation has a clear interpretation. The standardization processing of each indicator data eliminates the difference in the units of measurement between the data, making different indicators additive. This solution can choose to use the equal weight mode or the machine learning weighted mode according to actual needs, and can generate appropriate weights by combining it with the leave-one-out cross-validation method of the elastic network model. Based on this, the proposed solution in this embodiment can also obtain biological sample data at multiple time points, calculate the comprehensive assessment index at each time point, and perform trend analysis to predict the direction of disease progression, treatment effect and recurrence risk, thus realizing the leap from static assessment to dynamic prediction. It can also output visualized severity classification, radar chart and risk prediction report. The solution can also divide multiple target objects into several groups, calculate the average score of each group, and then realize the comparison results of group differences for evaluating the model.
[0016] Therefore, this invention proposes a method and system for assessing inflammatory bowel disease. The proposed solution integrates multiple indicator data to construct a multi-dimensional comprehensive assessment index, which can comprehensively and accurately reflect the pathophysiological nature of inflammatory bowel disease and effectively avoid misdiagnosis or missed diagnosis caused by single-dimensional assessment. It provides a non-invasive, convenient, and standardized technical solution for the diagnosis, disease monitoring, and efficacy evaluation of inflammatory bowel disease. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 A flowchart illustrating the overall process of the inflammatory bowel disease assessment method provided in Example 1; Figure 2 This is a schematic diagram of the methodology for assessing inflammatory bowel disease. Figure 3 This is a schematic diagram of a method for preprocessing biological sample data. Figure 4 A flowchart illustrating the method for training an elastic network model; Figure 5 A schematic diagram of the methodology for trend analysis of inflammatory bowel disease in target subjects; Figures 6(a) and 6(b) are both BIMI radar charts, where Figure 6(a) is an equal-weighted BIMI radar chart and Figure 6(b) is a machine learning-weighted BIMI radar chart. Figure 7 A flowchart illustrating the methodology for comparing group differences; Figures 8(a)-8(l) are comparative charts of methods for constructing comprehensive evaluation indices using equal weighting and machine learning weighting. Figure 8(a) is an example of equal-weighted BIMI, Figure 8(b) is an example of machine learning-weighted BIMI, Figure 8(c) is a correlation analysis chart of equal-weighted BIMI and DAI, Figure 8(d) is a correlation analysis chart of machine learning-weighted BIMI and DAI, Figure 8(e) is a receiver operating characteristic (ROC) curve of equal-weighted BIMI, and Figure 8(f) is a ROC curve of machine learning-weighted BIMI. (g) is an example of the correlation analysis between unstandardized equal-weighted BIMI and DAI; Figure 8(h) is an example of the correlation analysis between unstandardized machine learning weighted BIMI and DAI; Figure 8(i) is an example of the correlation analysis between unstandardized equal-weighted BIMI; Figure 8(j) is an example of the correlation analysis between unstandardized machine learning weighted BIMI; Figure 8(k) is an example of the SHAP analysis of small-sample machine learning weighted BIMI; Figure 8(l) is an example of the comparison results of each group of DAI. Figure 9 This is a schematic diagram of the module structure of the inflammatory bowel disease assessment system provided in Example 2. Detailed Implementation
[0019] 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 a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0020] Various embodiments of the invention will be described more fully below. The invention may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the invention to the specific embodiments disclosed herein, but rather the invention should be understood to cover all modifications, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of the invention.
[0021] In the following, the terms “comprising” or “may include” as used in various embodiments of the invention indicate the presence of the disclosed functions, operations, or elements, and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in various embodiments of the invention, the terms “comprising,” “having,” and their cognates are intended only to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or the possibility of adding one or more combinations of the foregoing.
[0022] In various embodiments of the invention, the expression "or" or "at least one of A and / or B" includes any combination or all combinations of the words listed simultaneously. For example, the expression "A or B" or "at least one of A and / or B" may include A, may include B, or may include both A and B.
[0023] The expressions used in the various embodiments of the present invention (such as "first," "second," etc.) may modify various constituent elements in the various embodiments, but do not limit the corresponding constituent elements. For example, the above expressions do not limit the order and / or importance of the elements. The above expressions are only used for the purpose of distinguishing one element from other elements. For example, a first user device and a second user device refer to different user devices, although both are user devices. For example, a first element may be referred to as a second element without departing from the scope of the various embodiments of the present invention, and similarly, a second element may also be referred to as a first element.
[0024] It should be noted that, in this invention, unless otherwise explicitly specified and defined, terms such as "installation," "connection," and "fixation" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0025] In this invention, those skilled in the art should understand that the terms indicating orientation or positional relationship in the text are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the purpose of facilitating the description of this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.
[0026] The terminology used in the various embodiments of the invention is for the purpose of describing particular embodiments only and is not intended to limit the various embodiments of the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. The terms (such as those defined in a generally used dictionary) are to be interpreted as having the same meaning as in the context of the relevant technical field and are not to be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.
[0027] Example 1 Please see Figures 1-2 This embodiment proposes a method for assessing inflammatory bowel disease. The proposed method can construct a dynamic risk prediction model, possessing dynamic risk prediction capabilities. Standardization ensures the comparability of results, and it simultaneously supports both static and dynamic assessment modes. It achieves comprehensive assessment consistent with disease progression without invasive procedures such as endoscopy. It can meet the needs of high-precision prediction in multiple scenarios and has advantages such as ease of operation, clinical applicability, and high detection accuracy. The method specifically includes: S100: Obtain biological sample data of the target object.
[0028] In this embodiment, the biological sample data includes at least barrier function index data, inflammatory factor index data, and gut microbiota index data; In a specific implementation using mice as the target, the method proposed in this embodiment uses existing mouse ulcerative colitis (UC) model static data as an example to construct a three-dimensional indicator library specifically for inflammatory bowel disease. Multiple core indicators were screened through a systematic literature review method, and a standardized indicator library was established according to functional classification to ensure coverage of the three core dimensions including epithelial barrier, inflammatory factors, and microorganisms. Among them, the barrier function index data includes protective data and damage data. The higher the expression value of the protective data, the more complete the barrier is, while the higher the expression value of the damage data, the more damaged the barrier is. Protective data may include, but are not limited to, data on MUC2 (mucin, maintaining the mucus layer), CLDN1 (tight junction protein, maintaining epithelial integrity), LAMC1 (laminusoidal protein, basement membrane structure), OLCN (closure protein, tight junction integrity), TJP1 (connector complex protein, anchoring tight junctions), CDH1 (epithelial cadherin, intercellular adhesion), MAGl1 (scaffold protein, regulating tight junctions), and HNF4A (transcription factor, maintaining epithelial cell function). Damage-related data may include, but are not limited to, data from RETNLB (resistance-like molecules that disrupt the mucus layer) and MEP1A (metalloproteinases that degrade the epithelial matrix).
[0029] Inflammatory factor indicators include pro-inflammatory and anti-inflammatory data. Higher expression values of pro-inflammatory data indicate a more severe degree of inflammation, while higher expression values of anti-inflammatory data indicate a milder degree of inflammation. Pro-inflammatory data may include, but are not limited to, data on TNF-α (a core pro-inflammatory factor), IL1B (an inflammation initiator), IFNG (a Th1-type inflammatory factor), IL23R (a Th17-type inflammatory receptor), IL6 (an inflammation amplifying factor), IL21 (a Th17-type cell activating factor), IL15 (a T-cell activating factor), IL17A (a Th17-type pro-inflammatory factor), CXCL1 (a neutrophil chemokine), S100A8 / S100A9 (calprotectin, an inflammation marker), and CCL2 (a monocyte chemoattractant protein). Anti-inflammatory data may include, but are not limited to, data on IL10 (anti-inflammatory factor, inhibiting pro-inflammatory factors), IL4 (Th2 factor, inhibiting Th1 inflammation), and TGFB1 (transforming growth factor, promoting tissue repair).
[0030] Gut microbiome data includes data on beneficial bacteria and harmful bacteria. A higher relative abundance of beneficial bacteria indicates a healthier gut microbiota, while a higher relative abundance of harmful bacteria indicates a more disrupted gut microbiota. Data on beneficial bacteria may include, but is not limited to, data on Roseburia (butyric acid-producing bacteria, which maintains the acidic environment of the gut), Ruminococcus (degrades dietary fiber and produces short-chain fatty acids), Akkermansia (mucus-degrading bacteria, which promotes barrier repair), Lactobacillus (lactic acid bacteria, which inhibits harmful bacteria), Bifidobacterium (Bifidobacterium, which regulates immunity), and Muribaculum (mouse-specific beneficial bacteria, which maintains gut microbiota balance). Harmful bacteria data may include, but are not limited to, data on Clostridium (toxin-producing bacteria) and Enterococcus (opportunistic pathogens).
[0031] To address the issue of ambiguous meaning in existing indicators, this embodiment calculates the barrier function score by subtracting the mean z-score of the damage data from the mean z-score of the protective data; the inflammation score by subtracting the mean z-score of the pro-inflammatory data from the mean z-score of the anti-inflammatory data; and the gut microbiota score by subtracting the mean z-score of the harmful bacteria data from the mean z-score of the beneficial bacteria data. In other words, the microbiota score = mean z-score of beneficial bacteria - mean z-score of harmful bacteria. Therefore, the method proposed in this embodiment ensures that for all indicators, a higher value indicates better biological function and a milder disease severity.
[0032] In another optional implementation, the barrier function index data may also include data on other barrier-related genes such as Occludin and Claudin homologs, the inflammatory factor index data may also include data on other pro-inflammatory factors and anti-inflammatory factors, and the gut microbiota index data may also include data on other bacterial genera and additional bacterial community functional gene index data.
[0033] S200: Preprocess biological sample data to standardize the data of each indicator.
[0034] S300: Based on preprocessed biological sample data, calculate barrier function score, inflammation score, and gut microbiota score to characterize the individual state of the target object.
[0035] S400: A comprehensive assessment index is constructed based on barrier function score, inflammation score, and gut microbiota score to quantitatively assess inflammatory bowel disease in target subjects.
[0036] Specifically, such as Figure 3 As shown, the method for preprocessing biological sample data in step S200 includes: S210: Transform the microbial abundance data in the gut microbiota index data.
[0037] Preferably, the abundance of microbial genera in step S210 is converted to a central log-ratio (CLR) transformation, i.e. ,in Let represent the abundance of the i-th bacterial genus, and D represent the number of bacteria included in the calculation in a single sample. Let be the abundance of the j-th genus.
[0038] S220: Convert gene FPKM values in barrier function index data and inflammatory factor index data.
[0039] Preferably, the transformation in step S220 may include log2 transformation, log10 transformation, or Box-Cox transformation; exemplarily, the method proposed in this embodiment can use a formula to process the FPKM values of 10 barrier function genes and 15 inflammatory factor genes. Transformation, in which, The FPKM value of the representative gene is... Increasing the value by 1 can avoid When the value is 0, taking the logarithm is meaningless.
[0040] It should be noted that by performing central logarithmic ratio transformation on the microbial abundance data in step S210 and transforming the gene FPKM value in step S220, the method proposed in this embodiment can solve the problems of the logarithmic distribution of microbial abundance and gene FPKM value and the large influence of extreme values, so that the data meets the requirements of statistical analysis.
[0041] S230: Standardize the data for each indicator to make different indicator data additive.
[0042] Preferably, the standardization process in step S230 includes Z-Score standardization or max-min standardization. In this embodiment, the standardization process includes Z-Score standardization. Z-Score standardization can standardize the three major categories of indicators—barrier, inflammation, and microorganism—by individual dimensions. The formula for Z-Score standardization is:
[0043] in, This represents the standardized sample or individual values. Represents the original sample value or individual value. This represents the sample mean of all samples or all individuals for that indicator. The standard deviation can eliminate the dimensional differences of each indicator, ensure the additivity of different indicators, and thus achieve comparability across indicators.
[0044] In this embodiment, the method for constructing the comprehensive evaluation index in step S400 includes determining the average value of the barrier function score, inflammation score, and gut microbiota score as the comprehensive evaluation index; or, by completing the training of the elastic network model, determining the weight coefficients for the barrier function score, inflammation score, and microbiota score respectively, and determining the weighted calculation result of the barrier function score, inflammation score, and gut microbiota score as the comprehensive evaluation index.
[0045] In one specific implementation, the formula for determining the comprehensive evaluation index using the mean may include: Barrier_Score = Mean(Z) MUC2 Z CLDN1 Z LAMC1 Z OCLN Z TJP1 Z CDH1 Z MAGl1 Z HNF4A )-Mean(Z RETNLB Z MEP1A ); Inflammation_Score = Mean(Z) TNF-α Z IL1B Z IFNG Z IL23R Z IL6 Z IL21 Z IL15 Z IL17A Z CXCL1 Z S100A8 Z S100A9 Z CCL2 )-Mean(Z IL10 Z IL4 Z TGFB1 ); Microorganism Score = Mean(Z) Roseburia Z Ruminococcus Z Akkermansia Z Lactobacillus Z Bifidobacterium Z Muribaculum )-Mean(Z Clostridium Z Enterococcus ); BIMI (Comprehensive Assessment Index) = (Barrier_Score + Inflammation_Score + Microorganism_Score) / 3.
[0046] like Figure 4 As shown, the specific methods for training elastic network models include: S410: Obtain a training dataset containing barrier function scores, inflammation scores, gut microbiota scores, and corresponding disease activity scores, Mayo scores, SES-CD scores, and / or CDEIS scores for multiple individual samples.
[0047] S420: Using barrier function score, inflammation score, and gut microbiota score from the training dataset as input features, and disease activity score, Mayo score, SES-CD score, and / or CDEIS score as prediction targets, construct a resilient network model.
[0048] It should be noted that the method proposed in this example can use any one or more of the disease activity score, Mayo score, SES-CD score, and CDEIS score as the prediction target, and can also use the Mayo score, SES-CD score, and CDEIS score as dependent variables while using the disease activity score as the prediction target.
[0049] S430: The elastic network model is trained using leave-one-out cross-validation to optimize the mixing parameters and regularization strength of the elastic network model in order to minimize the model prediction error until the preset convergence condition is met, thus completing the training of the elastic network model.
[0050] In an optional implementation, the method proposed in this embodiment can optimize weights based on the ElasticNet algorithm combined with leave-one-out cross-validation (LOOCV). Using the DAI (Disease Activity Index) as the dependent variable and the scores of the three major function categories as independent variables, an ElasticNet model (α=0.05) is constructed. The stability of the weights is verified through 100 bootstrap samplings, and a 95% confidence interval is calculated. The optimal weights are obtained through the trained ElasticNet model, where the weight of Barrier_Score is ω1, the weight of Inflammation_Score is ω2, and the weight of Microorganism_Score is ω3. The specific formula for determining the comprehensive evaluation index may include: BIMI=ω1 Barrier_Score+ω2 Inflammation_Score+ω3 Microorganism_Score Based on this, the method proposed in this embodiment can also verify feature importance through SHAP (SHapley Additive ex Planations). By analyzing the average absolute SHAP values of the three major categories, the biological rationale of the weights can be verified, ensuring that the results are interpretable.
[0051] In other embodiments, the method proposed in this embodiment can replace the ElasticNet regression model with algorithms such as Random Forest (RF), Support Vector Machine (SVM), and Gradient Boosting Tree (XGBoost) to calculate the weights of each dimension. Alternatively, the Delphi method can be used to determine the weights to suit situations where gold standard data is lacking.
[0052] like Figure 5 As shown, the method proposed in this embodiment further includes: S510: Obtain biological sample data of the target object at multiple different time points.
[0053] S520: A comprehensive evaluation index is obtained based on the data of each biological sample at each corresponding time point.
[0054] Specifically, step S510 can obtain biological sample data in the same way as step S100, and step S520 can obtain the comprehensive evaluation index corresponding to each time point in the same way as steps S200-S400.
[0055] S530: Based on trend analysis of various comprehensive assessment indices, predict the progression of inflammatory bowel disease, treatment efficacy, and / or relapse risk in target subjects.
[0056] Preferably, step S530 can collect three major categories of indicators in transcriptomics and 16S rRNA at multiple time points for the same target subject, and calculate the comprehensive assessment index for each individual at each time point according to steps S100-S400. The trend of the comprehensive assessment index is analyzed by time series fitting methods (such as linear regression or LSTM model) to predict disease progression, treatment effect and relapse risk.
[0057] For example, when the comprehensive assessment index continues to rise, step S530 will predict that the condition has been relieved and the treatment is effective; if the comprehensive assessment index continues to decline, step S530 will predict that the condition has worsened and the treatment plan needs to be adjusted; if the fluctuation range of the comprehensive assessment index is greater than the preset threshold, step S530 will predict that the condition is unstable and indicate that close monitoring is needed.
[0058] Optionally, the method proposed in this embodiment can also use moving average or exponential smoothing to replace linear trend fitting, improve the flexibility of dynamic prediction, and adjust the BIMI fluctuation threshold according to clinical data of different specific populations such as children and the elderly to obtain targeted treatment plans that meet the needs of different specific populations.
[0059] Please refer to Figures 6(a)-6(b). Based on this, the method proposed in this embodiment also includes generating assessment results after constructing a comprehensive assessment index. The assessment results include the barrier function score, inflammation score, microbial score and comprehensive assessment index corresponding to a single time point, as well as the severity grade of inflammatory bowel disease corresponding to the comprehensive assessment index. And / or, display radar charts showing the barrier function score, inflammation score, microbial score, and comprehensive assessment index for a single time point; And / or, a dynamic risk prediction report generated based on a comprehensive assessment index corresponding to multiple time points, including disease trends, probability of worsening, probability of remission, and / or treatment recommendations.
[0060] like Figure 7 As shown, preferably, the method proposed in this embodiment further includes: S610: Divide multiple different target objects into several groups.
[0061] S620: Based on the barrier function score, inflammation score, and gut microbiota score of all target subjects in each group, calculate the mean barrier function score, mean inflammation score, and mean gut microbiota score of each group to achieve comparison of population differences between different groups.
[0062] It should be noted that comparing group differences between different groups can enable model evaluation, helping users select and optimize the models they use.
[0063] For example, in step S610, multiple different target subjects can be divided into a control group, a disease group, and a treatment group, and the scores calculated in step S620 may include: Control_Barrier_Score (Control group barrier score) = mean(A) i _Barrier_Score); Control_Inflammation_Score (Inflammation score of the control group) = mean(A i (_Inflammation_Score); Control_Microorganism_Score (Control group microbial score) = mean(A i (Microorganism Score). DSS_Barrier_Score (DSS group barrier score) = mean(B i _Barrier_Score); DSS_Inflammation_Score (DSS group inflammation score) = mean(B i (_Inflammation_Score); DSS_Microorganism_Score (DSS group microbial score) = mean(B i (Microorganism Score). Treatment_Barrier_Score (Treatment group barrier score) = mean(C i _Barrier_Score); Treatment_Inflammation_Score (Inflammation Score of Treatment Group) = mean(C i (_Inflammation_Score); Treatment_Microorganism_Score (treatment group microbial score) = mean(C i _Microorganism_Score).
[0064] The method proposed in this embodiment has been validated in a mouse UC model, which solves the problem of the current technology being one-sided in terms of dimensions. It can comprehensively capture the progression of the disease, avoid misjudgment from a single dimension, and has the advantages of high accuracy and good predictive performance. The results are comparable through standardization. It is also non-invasive, convenient, easy to operate, and easy to translate into clinical practice. It can be applied to dynamic risk monitoring and meets the needs of high-precision prediction in multiple scenarios.
[0065] Please refer to Figures 8(a)-8(b). To verify the efficacy of the BIMI index proposed in this invention for assessing inflammatory bowel disease, a control group (Ctr), a disease group (DSS group), and a treatment group (T+1d group and T+4d group) were set up in a specific embodiment for inter-group comparison. The experimental results showed that there were significant differences between equal-weighted BIMI and machine learning-weighted BIMI in different groups (equal-weighted BIMI: P < 0.01; weighted BIMI: P < 0.001). However, a single indicator among barrier function, inflammatory factors, or microbial indicators can only distinguish some groups and cannot fully reflect the disease status. This confirms that the method proposed in this embodiment, by integrating the comprehensive assessment index constructed by the three dimensions, can more comprehensively capture the pathophysiological changes of IBD and is significantly better than the single-dimensional assessment method. Please refer to Figures 8(c)-8(f). In this embodiment, the accuracy of the BIMI index was verified in a mouse UC model using DAI as the reference standard. The verification results show that the Spearman correlation coefficient between equal-weighted BIMI and DAI is -0.747 (P < 0.0001), while the correlation coefficient between machine learning-weighted BIMI and DAI is improved to -0.765 (P < 0.0001), which is significantly better than the correlation range of a single indicator (r = -0.7554~0.67). The results indicate that the consistency between the BIMI index and the clinical gold standard is significantly improved after optimizing the weights through machine learning. Using DAI > 1 as the criterion for disease positivity, the AUC value of equal-weighted BIMI predicting the disease is 0.782, which has a moderately high diagnostic accuracy. The AUC value of machine learning-weighted BIMI is as high as 0.865, which is close to the perfect prediction level, indicating that the weighted BIMI index proposed in this invention has excellent diagnostic efficacy for inflammatory bowel disease. Please refer to Figures 8(g)-8(h). The method proposed in this embodiment eliminates the differences in the dimensions of indicators and the confusion in direction through the central logarithmic ratio transformation (CLR), log2 transformation, biological equilibrium algorithm, and Z-Score standardization process. It solves the problems of non-standard technical process and incomparable results at the current stage. The Spearman correlation coefficient between the unstandardized equal-weighted BIMI and DAI score is -0.74 (P<0.0001), and the Spearman correlation coefficient between the unstandardized machine learning weighted BIMI and DAI score is -0.743 (P<0.0001). Both are smaller than the corresponding data after the standardization process. That is, the data after standardization can more accurately reflect the severity of inflammatory bowel disease. Please refer again to Figures 8(a)-8(b). By capturing the trend of disease progression using BIMI values at multiple time points, the BIMI value of the T+1d treatment group (0.3861±0.6073) was significantly higher than that of the DSS model group (-1.3420±0.6810, P<0.05), while the BIMI value of the T+4d treatment group (0.4506±0.4071, P<0.05) was significantly higher than that of the DSS model group (-1.34). The BIMI value of the T+1d treatment group was 20±0.6810, which was higher than that of the T+1d treatment group. In the machine learning weighted BIMI, the BIMI value of the T+1d treatment group (0.2635±0.3690) was significantly higher than that of the DSS modeling group (-0.9713±0.4419, P<0.05), while the BIMI value of the T+4d treatment group (0.1858±0.2372) was higher than that of the DSS modeling group (-0.9713±0.4419). Please refer to Tables 1-2. Curcumin is a natural compound extracted from turmeric. Due to its broad safety profile and various biological activities such as anti-inflammatory, antioxidant, antibacterial, and immunomodulatory effects, it has shown great potential in the treatment of ulcerative colitis (UC). Since UC mice may be in the inflammatory phase at T+4d, curcumin shows better therapeutic effects during this period. It also shows that the BIMI value increases with the extension of treatment time, successfully predicting treatment effectiveness and solving the problem that current assessment techniques cannot dynamically predict treatment effectiveness.
[0066] Table 1. Scores and Composite Index of the Three Major Categories under Equal Weighted BIMI
[0067] Table 2. Scores and Composite Index of the Three Categories under Machine Learning Weighted BIMI
[0068] It should be noted that equal-weighted BIMI is simple to operate and highly interpretable, making it suitable for rapid assessment in primary healthcare institutions, but it neglects the biological significance of the three dimensions; while machine learning-weighted BIMI is more accurate and suitable for scenarios such as precision treatment and drug efficacy evaluation.
[0069] Please refer to Figures 8(i)-8(j). After standardization, the BIMI correlation between the two weights is 0.904 (P < 0.0001), which is greater than the BIMI correlation between the two weights without standardization, which is 0.690 (P < 0.0001). Moreover, there are only 10 different severity levels, but only 2 of them span two levels. The accuracy with DAI reaches 61.54%, which can effectively ensure the consistency of results in different scenarios. Even under the small sample conditions shown in Figure 8(k), after rigorous cross-validation, the model can still achieve a moderate effect and the prediction error is less than 1 DAI unit, which proves the robustness of the method proposed in this embodiment. Please refer to Figure 8(k). The SHAP analysis results show that the average contribution of microorganisms to the model output is 68.87%, which is higher than that of inflammatory factors (25.47%) and barrier function (5.66%). This result indicates that gut microbiota is the dominant factor driving changes in the BIMI index, which is highly consistent with the existing understanding that "microorganisms are not only triggers of inflammation, but also regulators of the barrier and regulators of immunity" in the pathogenesis of inflammatory bowel disease.
[0070] In the method proposed in this embodiment, the index detection can be carried out by conventional experimental methods such as qPCR gene expression detection and 16S rRNA sequencing. The BIMI value can be automatically generated by software, and the intuitive results can be output to facilitate rapid decision-making by clinicians.
[0071] It should be noted that although this embodiment provides validation data based on a mouse UC model, it does not represent a limitation on the method proposed in this embodiment. In practical applications, the method proposed in this embodiment can also be applied to the evaluation of inflammatory bowel diseases in other organisms, such as the clinical evaluation of inflammatory bowel diseases such as ulcerative colitis and Crohn's disease in humans.
[0072] Example 2 Please see Figure 9 This embodiment proposes an inflammatory bowel disease assessment system for implementing the method proposed in Embodiment 1. The system includes: The acquisition module 10 is used to acquire biological sample data of the target object; the biological sample data includes at least barrier function indicator data, inflammatory factor indicator data, and gut microbiota indicator data; Processing module 20 is used to preprocess biological sample data to eliminate dimensional differences between various indicator data; The calculation module 30 is used to calculate the barrier function score, inflammation score, and gut microbiota score, which characterize the individual state of the target object, based on the preprocessed biological sample data. Assessment module 40 is used to construct a comprehensive assessment index for quantitatively assessing inflammatory bowel disease in target subjects based on barrier function score, inflammation score, and gut microbiota score.
[0073] Furthermore, the acquisition module 10 can acquire biological sample data of the target object at multiple different time points, and after processing by the processing module 20 and calculation module 30, perform preset steps, ultimately enabling the evaluation module 40 to obtain a comprehensive evaluation index corresponding to each time point based on the biological sample data. The system also includes: Analysis module 50 is used to perform trend analysis based on the comprehensive assessment index at each corresponding time point to predict the disease progression, treatment effect and / or recurrence risk of inflammatory bowel disease in the target subjects.
[0074] Preferably, the system proposed in this embodiment further includes: The segmentation module 60 is used to divide multiple different target objects into several groups; The comparison module 70 is used to calculate the average barrier function score, average inflammation score, and average gut microbiota score of each group based on the barrier function score, inflammation score, and gut microbiota score of all target subjects in each group, so as to realize the comparison of population differences between different groups.
[0075] In summary, this invention proposes a method and system for assessing inflammatory bowel disease. The proposed scheme integrates multiple indicator data to construct a multi-dimensional comprehensive assessment index, which can comprehensively and accurately reflect the pathophysiological nature of inflammatory bowel disease. It effectively avoids misdiagnosis or missed diagnosis caused by single-dimensional assessment, and provides a non-invasive, convenient, and standardized technical solution for the diagnosis, disease monitoring, and efficacy evaluation of inflammatory bowel disease.
[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for assessing inflammatory bowel disease, characterized in that, The method includes: Obtain biological sample data of the target object; the biological sample data includes at least barrier function index data, inflammatory factor index data, and gut microbiota index data. The biological sample data are preprocessed to standardize the indicator data. Based on the preprocessed biological sample data, the barrier function score, inflammation score, and gut microbiota score, which characterize the individual state of the target object, are calculated. Based on the barrier function score, the inflammation score, and the gut microbiota score, a comprehensive assessment index is constructed to quantitatively evaluate the inflammatory bowel disease of the target subjects.
2. The method for assessing inflammatory bowel disease according to claim 1, characterized in that, The method for preprocessing the biological sample data includes: The microbial abundance data in the gut microbiota index data is transformed; The gene FPKM values in the barrier function index data and the inflammatory factor index data are converted; Each of the aforementioned indicator data is standardized to make the different indicator data additive.
3. The method for assessing inflammatory bowel disease according to claim 2, characterized in that, The transformation performed on the relative abundance data of the microorganisms includes central log ratio transformation; The transformations performed on the gene FPKM values include log2 transformation, log10 transformation, or Box-Cox transformation; The standardization process includes Z-Score standardization or max-min standardization.
4. The method for assessing inflammatory bowel disease according to claim 1, characterized in that, The barrier function index data includes at least protective data and damaging data, and the barrier function score is obtained by subtracting the mean z-score of the damaging data from the mean z-score of the protective data. The inflammatory factor index data includes at least pro-inflammatory data and anti-inflammatory data, and the inflammation score is obtained by subtracting the mean z-score of the pro-inflammatory data from the mean z-score of the anti-inflammatory data. The gut microbiota index data includes at least beneficial bacteria data and harmful bacteria data, and the gut microbiota score is obtained by subtracting the mean z-score of the harmful bacteria data from the mean z-score of the beneficial bacteria data.
5. The method for assessing inflammatory bowel disease according to claim 1 or 4, characterized in that, The method for constructing the comprehensive evaluation index includes: The average of the barrier function score, the inflammation score, and the gut microbiota score is determined as the comprehensive evaluation index; Alternatively, by training an elastic network model, weight coefficients can be determined for the barrier function score, inflammation score, and gut microbiota score, and the weighted calculation result of the barrier function score, inflammation score, and gut microbiota score can be determined as the comprehensive evaluation index.
6. The method for assessing inflammatory bowel disease according to claim 5, characterized in that, The method for training the elastic network model includes: Obtain a training dataset containing barrier function scores, inflammation scores, gut microbiota scores, and corresponding disease activity scores, Mayo scores, SES-CD scores, and / or CDEIS scores for multiple individual samples; The elastic network model is constructed by using barrier function score, inflammation score and gut microbiota score from the training dataset as input features, and disease activity score, Mayo score, SES-CD score and / or CDEIS score as prediction targets. The elastic network model is trained using leave-one-out cross-validation to optimize the mixing parameters and regularization strength of the elastic network model in order to minimize the model prediction error until the preset convergence condition is met, thus completing the training of the elastic network model.
7. The method for assessing inflammatory bowel disease according to claim 1, characterized in that, The method further includes: Obtain biological sample data of the target object at multiple different time points; A comprehensive evaluation index corresponding to each of the aforementioned time points is obtained based on the biological sample data. Trend analysis is performed based on the comprehensive assessment indices to predict the progression of inflammatory bowel disease, treatment efficacy, and / or recurrence risk in the target subjects.
8. The method for assessing inflammatory bowel disease according to claim 7, characterized in that, The method further includes generating an evaluation result after constructing the comprehensive evaluation index, the evaluation result including: The barrier function score, inflammation score, microbial score, and comprehensive assessment index at a single time point, as well as the severity grading of inflammatory bowel disease corresponding to the comprehensive assessment index; And / or, display radar charts showing the barrier function score, inflammation score, microbial score, and comprehensive assessment index for a single time point; And / or, a dynamic risk prediction report generated based on a comprehensive assessment index corresponding to multiple time points, including disease trends, probability of worsening, probability of remission, and / or treatment recommendations.
9. The method for assessing inflammatory bowel disease according to claim 1, characterized in that, The method further includes: Divide multiple different target objects into several groups; Based on the barrier function score, inflammation score, and gut microbiota score of all target subjects in each group, the average barrier function score, average inflammation score, and average gut microbiota score of each group are calculated to achieve a comparison of population differences between different groups.
10. An assessment system for inflammatory bowel disease, characterized in that, The system includes: The acquisition module is used to acquire biological sample data of the target object; the biological sample data includes at least barrier function index data, inflammatory factor index data, and gut microbiota index data. The processing module is used to preprocess the biological sample data to eliminate dimensional differences between the various indicator data; The calculation module is used to calculate the barrier function score, inflammation score, and gut microbiota score, which characterize the individual state of the target object, based on the preprocessed biological sample data. The assessment module is used to construct a comprehensive assessment index for quantitatively assessing the inflammatory bowel disease of the target subject based on the barrier function score, the inflammation score, and the gut microbiota score.