informational program evaluation system for therapeutic effect of active ulcerative colitis

Abstract

This application discloses an information-based planning and evaluation system for the treatment of active ulcerative colitis, belonging to the field of planning and evaluation technology. By analyzing the state data of food and drugs after they pass through the digestive tract, and calculating sharpness and residue load, it quantifies the potential mechanical stimulation or burden that different foods and drugs may cause to the fragile and damaged colonic mucosa. It provides a more sensitive and accurate quantitative tool by automatically extracting microscopic feature data from endoscopic images. In addition to predicting drug efficacy, it also predicts mucosal healing trends under specific food patterns. Regarding drug recommendations, it obtains a drug recommendation list through weighted integration using machine learning models and the aforementioned prediction results. For dietary advice, it combines association rule mining and Bayesian networks to obtain specific food pattern suggestions, improving the targeting and patient suitability of dietary management, and enhancing the transparency and safety of clinical decision-making.

Description

Technical Field

[0001] This application falls under the field of planning evaluation, specifically an information-based planning evaluation system for the treatment efficacy of active ulcerative colitis. Background Technology

[0002] Ulcerative colitis is a chronic, relapsing inflammatory bowel disease characterized by diffuse inflammation, erosion, and ulceration of the colonic mucosa during active phases. Its treatment and management are highly complex and individualized. Current clinical practice primarily relies on disease activity indices, endoscopic assessments, and patient-reported symptoms to develop treatment plans, including medication selection and dietary recommendations. However, existing methods have significant limitations: First, treatment decisions rely on experience and lack detailed and quantitative analysis of how drugs and food specifically affect the mucosal healing process; in particular, dietary management is mostly based on general principles (such as low-residue diets) and cannot predict the potential mechanical stimulation or impact on the damaged mucosa of a specific patient based on the physical state of food as it passes through the digestive tract (such as the sharpness of the residue after the hardness decreases and the fiber load). Secondly, the efficacy prediction and evaluation are based on a single dimension, mainly relying on the subjective visual judgment of endoscopists. There is a lack of objective and quantitative data extraction and analysis for subtle but crucial healing indicators such as the clarity of mucosal vascular texture and branching morphology. Finally, the system's ability to generate treatment plans is insufficient. Existing recommendation systems often separate medication and dietary advice, failing to deeply integrate patients' medication history, individual treatment response, dietary logs, and objective predictions of mucosal healing to construct dynamic, evidence-based management plans. To address the problems raised in this background, this application designs an information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis. Summary of the Invention

[0003] To address the aforementioned technical shortcomings, this application proposes an information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis.

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: This application provides an information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, which includes the following specific modules: The colitis data acquisition module is used to acquire data on the state of various foods and drugs after they pass through the digestive tract, patient endoscopic image data, mucosal damage data, and patient characteristic data. The Recovery Impact Analysis module is used to analyze the recovery impact of various foods and drugs based on the state data of various foods and drugs after passing through the digestive tract. The healing prediction and analysis module is used to predict and analyze the healing status of patients' lesions based on the results of the analysis of the impact of various foods and drugs on healing, patient endoscopic image data, and mucosal damage data. The patient food and drug matching analysis module is used to establish a patient food and drug matching analysis model, and to make recommendations on patient food and drugs based on the patient food and drug matching analysis model, the prediction analysis results of patient healing status corresponding to various foods and drugs, and patient characteristic data. The patient food and drug recommendation module is used to obtain a list of recommended medications and specific food patterns for patients, and then push the list of recommended medications and specific food patterns to relevant personnel for processing.

[0005] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the specific process for acquiring data on the state of various foods and drugs after passing through the digestive tract, patient endoscopic image data, mucosal damage data, and patient characteristic data is as follows: Data on the state of various foods and drugs after passing through the digestive tract is acquired through quantitative analysis results from drug laboratory tests, food databases, Micro-CT scans, and in vitro simulated digestion experiments. This data includes data on the hardness reduction of various foods and drugs after passing through the digestive tract, shape and contour image data, insoluble dietary fiber content data, and indigestible particulate matter content data. Endoscopic image data of the patient is acquired through image processing operations on the patient's endoscopic images. According to the data, the patient's endoscopic image data includes contrast data of blood vessels in the endoscopic images, the ratio of blood vessel pixels to total mucosal pixels, and the number of blood vessel branches. Mucosal damage data is obtained through image processing and OTC measurements of the mucosal images, including the area and depth of damage in each damaged region. Patient characteristic data is obtained through the hospital information system, including patient characteristic data, food model data, and drug model data. Patient characteristic data includes age, gender, BMI, disease duration, surgical history, treatment history, historical medications, and daily dietary habits. Drug model data includes drug type, name, dosage, and duration of use. Food model data includes daily dietary habits, food intake, and frequency of meals.

[0006] It should be noted that, as the preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the analysis process for the impact of various foods and drugs on recovery is as follows: The sharpness analysis results of various foods and drugs were obtained by using data on the hardness decay of various foods and drugs after passing through the digestive tract and shape contour image data. The analysis results of the residue load of various foods and drugs were obtained by using data on the content of insoluble dietary fiber and undigested particulate matter after various foods and drugs passed through the digestive tract. Based on the analysis results of the sharpness and residue load of various foods and medicines, the healing impact of various foods and medicines was analyzed.

[0007] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the process of obtaining the sharpness analysis results of various foods and drugs from the data on the attenuation of their hardness and the image data of their shape contours after passing through the digestive tract includes the following specific steps: Analyzing the sharpness of various foods and drugs based on the data on the attenuation of their hardness and the image data of their shape contours after passing through the digestive tract, wherein the formula for analyzing the sharpness of the i-th type of food and drug is: Let i be the number corresponding to each food and medicine, where i can be any number from 1 to N, and N be the maximum number corresponding to each food and medicine. Let i be the rate at which the hardness of the i-th type of food and medicine decreases. Let be the shape parameters of the hardness decay curve for the i-th type of food and medicine. For reference to the softening half-life, Let be the minimum radius of curvature corresponding to the shape contour line of the i-th type of food and medicine. and These are the weights corresponding to the hardness decay and the minimum radius of curvature, respectively.

[0008] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the efficacy of active ulcerative colitis, the analysis results of the residue load of various foods and drugs obtained from the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after various foods and drugs pass through the digestive tract include the following specific steps: The residue load analysis of various foods and drugs is performed based on the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after various foods and drugs pass through the digestive tract. The process of analyzing the residue load of various foods and drugs is as follows: the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after various foods and drugs pass through the digestive tract are divided by the reference content of insoluble dietary fiber and the content of indigestible particulate matter, respectively, to obtain the food residue load and the drug residue load. The weighted sum of the food residue load and the drug residue load yields the analysis results of the residue load of various foods and drugs.

[0009] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the analysis of the impact of various foods and drugs on healing includes the following specific steps: Based on the results of the sharpness analysis and residue load analysis of various foods and drugs, the analysis of the impact of various foods and drugs on healing is conducted. Specifically, the process of this analysis involves: obtaining the results of the sharpness analysis and residue load analysis of various foods and drugs; and then weighting and summing these results to obtain the final analysis result of the impact of various foods and drugs on healing.

[0010] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the specific process of predicting and analyzing the healing status of patient lesions corresponding to various foods and drugs is as follows: Obtaining the healing impact analysis results of various foods and drugs, contrast data of blood vessel images in the patient's endoscopic images, the ratio of blood vessel pixels to total mucosal pixels, the number of blood vessel branches, and the area and depth of each mucosal lesion. Based on the healing impact analysis results of various foods and drugs, the clarity of blood vessel texture in the patient's endoscopic images, and the mucosal lesion data, performing predictive analysis of the healing status of patient lesions corresponding to various foods and drugs. Specifically, the predictive analysis process for the healing status of patient lesions corresponding to various foods and drugs involves: processing the contrast data of blood vessel images, the ratio of blood vessel pixels to total mucosal pixels, and the number of blood vessel branches in the patient's endoscopic images separately... Normalization was performed, and the processed results were weighted and summed to obtain the clarity of the patient's blood vessels. The area of ​​each damaged mucosa region was summed to obtain the total area of ​​the damaged mucosa region. The total area of ​​the damaged mucosa region was divided by the total mucosa area to obtain the proportion of mucosal damage. The average depth of the damaged mucosa region was calculated by averaging the data of the damaged depth of each damaged mucosa region. The average depth of the damaged mucosa region was divided by the maximum depth of each damaged mucosa region to obtain the depth of the damaged mucosa region. The depth of the damaged mucosa region and the proportion of damaged mucosa region were added to obtain the degree of mucosal damage. The reciprocal of the clarity of the blood vessels was added to the degree of mucosal damage to obtain the severity of the patient's lesion. The severity of the patient's lesion was multiplied by the results of the analysis of the impact of various foods and drugs on healing to obtain the predictive analysis results of the healing of the patient's lesion corresponding to various foods and drugs.

[0011] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the specific process of establishing a patient food-medicine matching analysis model and recommending patient foods and medicines based on the patient food-medicine matching analysis model, the prediction analysis results of patient healing status corresponding to various foods and medicines, and patient characteristic data is as follows: The patient food-medicine matching analysis model includes a patient medicine recommendation sub-model and a food suggestion sub-model. The medicine recommendation sub-model has four layers: a patient medicine model data preprocessing layer, a basic prediction model layer for medicine recommendations, a basic prediction result integration layer, and a risk interpretation and warning layer; the food suggestion sub-model has four layers: The system comprises a patient food model data preprocessing layer, an association rule mining and Bayesian network layer, a diet pattern generation layer, and a diet plan construction layer. The drug recommendation sub-model construction process involves: processing patient feature data and drug model data in the patient-drug model data preprocessing layer to transform them into drug recommendation-related features; standardizing these features; constructing and training two basic random forest models in the basic prediction result ensemble layer; inputting the standardized drug recommendation-related features into the two basic random forest models to predict the duration of relief and the probability of relief for various drugs; and then... The results of the analysis of remission duration, remission probability, and patient healing status are weighted and averaged in the basic prediction result integration layer to obtain a comprehensive evaluation index for various drugs. Simultaneously, drugs are ranked according to their comprehensive evaluation indices to obtain a recommended drug list for each patient. In the risk warning explanation layer, explanations and risk warnings are added to the recommended drugs in the patient's drug list. The food suggestion sub-model construction process is as follows: patient feature data and food model data are processed for patient dietary structure in the patient food model data preprocessing layer to obtain food component decomposition results. In the association rule mining and Bayesian network layer, association rule mining is used to obtain various food components. The correlation between the decomposition results and the patient healing prediction analysis results is investigated. A Bayesian network is constructed to establish the conditional dependency between food components and the patient healing prediction analysis results. Based on the correlation between the decomposition results of various food components and the patient healing prediction analysis results, and the conditional dependency between these results, food component suggestions are generated for patients at the diet pattern generation layer. Based on these suggestions, specific food patterns are generated at the diet plan construction layer. The patient healing prediction analysis results and patient characteristic data corresponding to various foods and medications are input into the patient food-medicine matching analysis model to obtain a patient medication recommendation list and specific food patterns.

[0012] It should be noted that, as a preferred technical solution for the information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, the specific process of pushing the patient's recommended medication list and specific food patterns to relevant personnel for reference is as follows: merging the patient's recommended medication list and specific food patterns to generate a patient condition report, and then pushing the patient condition report to relevant personnel for processing; the patient condition report includes the recommended medication list, dietary advice, and precautions.

[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention acquires state data of various foods and medicines after they pass through the digestive tract, patient endoscopic image data, patient characteristic data, and mucosal damage data; it performs healing impact analysis on various foods and medicines based on the state data of various foods and medicines after they pass through the digestive tract; it acquires healing impact analysis results of various foods and medicines, contrast data of blood vessel images in patient endoscopic images, ratio data of blood vessel pixels to total mucosal pixels, data on the number of blood vessel branches, and data on the damage area and depth of each damaged area of ​​the mucosa; it predicts and analyzes the healing status of patient lesions corresponding to various foods and medicines based on the healing impact analysis results of various foods and medicines, data on the clarity of blood vessel texture in patient endoscopic images, and mucosal damage data; it establishes a patient food and medicine matching analysis model, and recommends patient foods and medicines based on the patient food and medicine matching analysis model, the patient healing status prediction analysis results corresponding to various foods and medicines, and patient characteristic data; it acquires a patient medicine recommendation list and specific food patterns, and integrates patient medicines... Recommended lists and specific food patterns are pushed to relevant personnel for processing; the state data of food and drugs after passing through the digestive tract are analyzed, and the mechanical stimulation or burden that different foods and drugs may cause to fragile and damaged colonic mucosa is quantified by calculating the sharpness and residue load; by automatically extracting microscopic features such as vascular contrast, vascular pixel ratio, and number of vascular branches in endoscopic images, a more sensitive and accurate quantitative tool is provided for the system; at the same time, not only is the efficacy of drugs predicted, but also the mucosal healing trend under specific food patterns; in terms of drug recommendations, the probability of relief and maintenance time are predicted by integrating machine learning models, and combined with the aforementioned healing predictions for weighted integration, a drug recommendation list sorted by a comprehensive evaluation index is generated, while providing evidence and risk descriptions for the recommendation results, enhancing the transparency and safety of clinical decision-making; in terms of dietary advice, specific food pattern suggestions are obtained by combining association rule mining and Bayesian networks. These suggestions are tailored to the patient's personal history and predicted healing trajectory, rather than a general diet, which greatly improves the pertinence and patient suitability of dietary management. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall framework of the information-based planning and evaluation system for the efficacy of ulcerative colitis during the active phase of this application.

[0015] Figure 2 This is a schematic diagram illustrating the impact analysis process of various foods and drugs on the cure of ulcerative colitis during the active phase of this application, which is part of the information-based planning and evaluation system for the efficacy of ulcerative colitis treatment.

[0016] Figure 3 This diagram illustrates the acquisition of various food and drug efficacy impact analysis results for the information-based planning and evaluation system for the treatment of ulcerative colitis during the active phase of this application.

[0017] Figure 4 This is a schematic diagram illustrating the specific process of predicting and analyzing the healing status of patient lesions corresponding to various foods and drugs in the information-based planning and evaluation system for the efficacy of ulcerative colitis during the active phase of this application.

[0018] Figure 5 This is a schematic diagram of the patient food and drug recommendation model for the information-based planning and evaluation system for the treatment efficacy of ulcerative colitis during the active phase of this application. Detailed Implementation

[0019] To better understand this application, various aspects of this application will be described in more detail with reference to the accompanying drawings.

[0020] To address the technical problems raised in the background art, this application provides a preferred embodiment: The specific content of this embodiment is as follows: like Figure 1 As shown, the information-based planning and evaluation system for the treatment of active ulcerative colitis includes the following specific modules: The colitis data acquisition module is used to acquire data on the state of various foods and drugs after they pass through the digestive tract, patient endoscopic image data, mucosal damage data, and patient characteristic data. In this embodiment, the specific process for obtaining state data of various foods and drugs after passing through the digestive tract, patient endoscopic image data, mucosal damage data, and patient characteristic data is as follows: State data of various foods and drugs after passing through the digestive tract includes data on the hardness decay, shape contour image data, insoluble dietary fiber content data, and indigestible particulate matter content data; the indigestible particulate matter content data is the percentage of undigested drug residue relative to the original drug mass, obtained through quantitative analysis results from a pharmaceutical laboratory; the insoluble dietary fiber content data is obtained from a food database; the shape contour image data of various foods and drugs after passing through the digestive tract is obtained through Micro-CT scanning of sampled samples; the hardness decay data of various foods and drugs after passing through the digestive tract includes the hardness decay rate and hardness decay curve shape parameters, obtained through in vitro simulated digestion experiments to obtain data points on the time and hardness changes of various foods and drugs, and analyzed through nonlinear regression analysis (e.g., maximum... The hardness decay rate and hardness decay curve shape parameters are obtained by solving the high likelihood estimation or least squares method; the patient endoscopic image data includes the contrast data of blood vessel images, the ratio of blood vessel pixels to total mucosal pixels, and the number of blood vessel branches in the patient endoscopic images, which are obtained through image processing operations on the patient endoscopic images; the mucosal damage data includes the damage area and depth data of each damaged area of ​​the mucosa, the damage depth of each damaged area of ​​the mucosa is obtained through OTC measurement, and the damage area of ​​each damaged area of ​​the mucosa is obtained through image processing of the mucosal images; the patient characteristic data includes patient characteristic data, food model data, and drug model data. The patient characteristic data includes age, gender, BMI, disease course, surgical history, treatment history, historical medication, and daily dietary habits, etc. The drug model data includes drug type, name, dosage, and duration of use, and the food model data includes daily dietary habits, food quantity, and frequency of meals, etc. The patient characteristic data, drug model data, and food model data are obtained through the hospital information system; The Recovery Impact Analysis module is used to analyze the recovery impact of various foods and drugs based on the state data of various foods and drugs after passing through the digestive tract. like Figure 2 As shown in this embodiment, the analysis process for the impact of various foods and medicines on recovery is as follows: The sharpness analysis results of various foods and drugs were obtained by using data on the hardness decay of various foods and drugs after passing through the digestive tract and shape contour image data. In this embodiment, obtaining the sharpness analysis results of various foods and drugs from the hardness attenuation data and shape contour image data after passing through the digestive tract includes the following specific steps: Analyzing the sharpness of various foods and drugs based on the hardness attenuation data and shape contour image data after passing through the digestive tract, wherein the sharpness analysis formula for the i-th type of food and drug is: Let i be the number corresponding to each food and medicine, where i can be any number from 1 to N, and N be the maximum number corresponding to each food and medicine. Let i be the rate at which the hardness of the i-th type of food and medicine decreases. Let be the shape parameters of the hardness decay curve for the i-th type of food and medicine. For reference to the softening half-life, Let be the minimum radius of curvature corresponding to the shape contour line of the i-th type of food and medicine. and These represent the weights corresponding to the hardness decay and the minimum radius of curvature, respectively; it should be noted that in this formula, The acquisition process is as follows: Shape contours are extracted from image data of the shape contours of various foods and medicines after they pass through the digestive tract. The curvature of each point along the extracted shape contour is calculated, and the radius of curvature is the reciprocal of the curvature, thus obtaining the minimum radius of curvature. The minimum radius of curvature is used to quantify the sharpness of its edges; the smaller the minimum radius of curvature, the sharper the edge. In this formula... Some results indicate the softening half-life of the i-th food and drug, i.e., the time required for the hardness of the i-th food and drug to decrease to half of its initial hardness value. This is derived from the definition of the Weibull model. The result of the Weibull model's definition formula is the hardness value of the food and drug at each digestion time, i.e., the hardness decay curve of the food and drug. When the shape parameter (which determines the shape of the hardness decay curve) of the i-th food and drug is less than 1, it indicates that the decay rate decreases with time. Specifically, the hardness decay curve of the food and drug decreases rapidly at the beginning and then gradually slows down, corresponding to the case where the surface of the food and drug quickly absorbs water and softens, but the internal dense structure decomposes slowly. When the shape parameter is equal to 1, it indicates that the decay rate is constant, corresponding to the case where the food and drug have a uniform structure and disintegrates at a constant rate. When the shape parameter is greater than 1, it indicates that the decay rate increases with time. At this time, the curve decreases slowly at the beginning and then accelerates disintegration after a certain point, corresponding to the case where many dense foods (such as beans and certain grains) require sufficient penetration of digestive juices and the enzyme action to reach a threshold before the structure rapidly disintegrates. The softening half-life... The larger the result or the slower the overall process of hardness decay (i.e.) The larger the value, the higher the risk of food and drugs remaining firm in the colonic segment, and the greater the risk of irritation to the ulcerated mucosa. The values ​​are determined by referring to the softening half-life, the degree of firmness decay, and the minimum radius of curvature. The experimental data were statistically analyzed, with the experimental data as the independent variable and the experimental results as the dependent variable, and multiple linear regression analysis was performed.

[0021] The analysis results of the residue load of various foods and drugs were obtained by using data on the content of insoluble dietary fiber and undigested particulate matter after various foods and drugs passed through the digestive tract. In this embodiment, the analysis results of the residue load of various foods and drugs obtained from the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after passing through the digestive tract include the following specific steps: The residue load analysis of various foods and drugs is performed based on the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after passing through the digestive tract. The process of analyzing the residue load of various foods and drugs is as follows: The data on the content of insoluble dietary fiber and the content of indigestible particulate matter after passing through the digestive tract of various foods and drugs are divided by a reference insoluble dietary fiber content. The fiber content and indigestible particulate matter content were used to obtain the food residue load and drug residue load, respectively. The weighted sum of these two loads yielded the analysis results for the residue load of various foods and drugs. It should be noted that the insoluble dietary fiber content data for food and drugs after passing through the digestive tract refers to the insoluble dietary fiber content per 100g of food and drug (e.g., approximately 4-5g of insoluble dietary fiber in 100g of whole wheat bread). Insoluble dietary fiber is insoluble in water, increases stool volume, and promotes intestinal peristalsis, which is beneficial to healthy patients. However, for patients with active ulcerative colitis, this is a physical burden. The insoluble dietary fiber content, or residue load (the total amount of substances that cannot be absorbed after food and drugs pass through the digestive tract, directly contact the inflamed intestinal mucosa, and may irritate it), and the data on undigested particulate matter content refer to the insoluble shells or carriers of some special dosage forms of drugs. These insoluble shells or carriers (such as sustained-release microspheres and capsule shells) exist as foreign bodies in the intestines, causing additional friction to the already inflamed and ulcerated intestinal walls. In cases of severe intestinal inflammation, intestinal stenosis, or abnormal intestinal motility, these insoluble shells or carriers can form drug-induced fecal matter. The potential risks of stones, and the fact that some drug carriers or ingredients (such as antacids or supplements containing aluminum, calcium, and iron) may alter the local intestinal environment, affecting the absorption of other drugs or directly irritating the mucosa; the weights of food residue load and drug residue load are the influence coefficients of food residue load and drug residue load on active ulcerative colitis obtained through data regression analysis; the reference insoluble dietary fiber content is the upper limit of daily insoluble dietary fiber intake for active ulcerative colitis, and the reference undigested particulate matter content is the average percentage of each undigested drug residue relative to the original drug mass.

[0022] Based on the analysis results of the sharpness and residue load of various foods and medicines, an analysis of the healing impact of various foods and medicines was conducted. like Figure 3As shown, in this embodiment, the analysis of the healing impact of various foods and drugs includes the following specific steps: The analysis of the healing impact of various foods and drugs is conducted based on the results of the sharpness analysis and residue load analysis. The process of this analysis involves obtaining the results of the sharpness analysis and residue load analysis of various foods and drugs, and then weighting and summing these results to obtain the final analysis results. It should be noted that the results of this analysis establish a link between the physical and chemical properties of food and drugs and the recovery process of active ulcerative colitis, integrating multidimensional information to analyze the impact on patient recovery. This considers both the sharpness of food and drugs and the volumetric burden of residue, providing data-driven support for patient health management and improving its scientific rigor.

[0023] The healing prediction and analysis module is used to predict and analyze the healing status of patients' lesions based on the results of the analysis of the impact of various foods and drugs on healing, patient endoscopic image data, and mucosal damage data. like Figure 4As shown, in this embodiment, the specific process of predicting and analyzing the healing status of patient lesions corresponding to various foods and drugs is as follows: Obtaining the healing impact analysis results of various foods and drugs, contrast data of blood vessel images in the patient's endoscopic images, the ratio of blood vessel pixels to total mucosal pixels, the number of blood vessel branches, and the area and depth of each damaged area of ​​the mucosa. Based on the healing impact analysis results of various foods and drugs, the clarity of blood vessel texture in the patient's endoscopic images, and the mucosal damage data, performing predictive analysis of the healing status of patient lesions corresponding to various foods and drugs. The predictive analysis process for the healing status of patient lesions corresponding to various foods and drugs involves: [The text abruptly ends here, likely due to an incomplete translation or source material.] The contrast data of blood vessels, the ratio of blood vessel pixels to total mucosal pixels, and the number of blood vessel branches in the endoscopic images were normalized (e.g., using max-min normalization or Z-score normalization). The processed results were then weighted and summed to obtain the clarity of the patient's blood vessels. The total area of ​​the damaged mucosa was obtained by summing the damaged areas of each region. The percentage of damaged mucosa was then calculated by dividing the total damaged area by the total mucosal area. The average depth of the damaged mucosa was obtained by averaging the depth of each damaged area. The maximum depth of the lesion is used to determine the depth of the patient's mucosal damage. Adding the depth of the mucosal damage to the percentage of damage results in the degree of mucosal damage. Adding the reciprocal of the clarity of the blood vessels to the degree of mucosal damage gives the severity of the affected area. Multiplying the severity of the affected area by the results of the impact analysis on healing from various foods and medications yields the predictive analysis results for the healing of the affected area corresponding to each food and medication. It should be noted that clear vascular texture is usually characterized by high contrast between the blood vessels (dark) and the surrounding mucosa (bright). When the blood vessels are blurred, the contrast decreases, meaning that contrast is positively correlated with vascular clarity. The ratio of the number of blood vessel pixels to the total number of mucosal pixels is inversely related. The algorithm reflects the proportion of visible blood vessels. When blood vessels are clear, the segmentation algorithm can extract the vascular network more completely, and the ratio of blood vessel pixels to total mucosal pixels is higher. When blood vessels are blurred or even disappear, the segmented vascular area decreases, and the ratio of blood vessel pixels to total mucosal pixels decreases. In other words, the ratio of blood vessel pixels to total mucosal pixels is positively correlated with the clarity of blood vessels. The number of blood vessel branches is used because a clear vascular network has abundant fine branches. When blood vessels are blurred, the fine branches disappear first, leading to a decrease in the number of detectable branches. In other words, the number of blood vessel branches is positively correlated with the clarity of blood vessels. The more and deeper the mucosal damage area, the more severe the damage (i.e., ulcer).

[0024] The patient food and drug matching analysis module is used to establish a patient food and drug matching analysis model, and to make recommendations on patient food and drugs based on the patient food and drug matching analysis model, the prediction analysis results of patient healing status corresponding to various foods and drugs, and patient characteristic data. like Figure 5As shown, in this embodiment, a patient food and drug matching analysis model is established. The specific process for recommending patient food and drugs based on this model, the prediction analysis results of patient healing status corresponding to various foods and drugs, and patient characteristic data is as follows: The patient food and drug matching analysis model includes a patient drug recommendation sub-model and a food suggestion sub-model. The drug recommendation sub-model has four layers: a patient drug model data preprocessing layer, a drug recommendation basic prediction model layer, a basic prediction result integration layer, and a risk explanation layer. The food suggestion sub-model also has four layers: a patient food model data preprocessing layer, an association rule mining and Bayesian network layer, a diet pattern generation layer, and a diet... The scheme construction layer includes the following process for constructing the drug recommendation sub-model: Patient feature data and drug model data are processed in the patient-drug model data preprocessing layer, transforming them into drug recommendation-related features. These features are then standardized. Two basic random forest models are constructed and trained in the basic prediction result integration layer. The standardized drug recommendation-related features are input into these two models to predict the duration and probability of relief for various drugs. Finally, the predicted duration and probability of relief for various drugs, along with the patient's healing outcome predictions, are weighted and averaged in the basic prediction result integration layer. The process involves obtaining comprehensive evaluation indices for various medications, ranking them according to these indices, and generating a recommended medication list for each patient. In the risk warning layer, explanations and risk warnings are added to the medications in this recommended list. The food suggestion sub-model construction process involves: preprocessing patient feature data and food model data in the patient food model data preprocessing layer to obtain food component decomposition results; then, in the association rule mining and Bayesian network layer, using association rule mining (Apriori algorithm) to obtain the correlation between the food component decomposition results and the patient's healing prediction analysis results; and simultaneously constructing a Bayesian network to establish the correlation between food components and patient... The conditional dependencies of the patient's healing status prediction analysis results are analyzed. Based on the correlation between the decomposition results of various food components and the patient's healing status prediction analysis results, and the conditional dependencies between food components and the patient's healing status prediction analysis results, food component suggestions are generated for the patient at the diet pattern generation layer. Based on the patient's food component suggestions, specific food patterns are generated at the diet plan construction layer. The patient's healing status prediction analysis results and patient characteristic data corresponding to various foods and drugs are input into the patient's food and drug matching analysis model to obtain the patient's drug recommendation list and specific food patterns. It should be noted that the drug recommendation sub-model and the food suggestion sub-model should be trained in parallel. The model training process is existing technology and will not be described in detail here.In constructing the drug recommendation sub-model and food suggestion sub-model, during data processing at the data preprocessing layer, it is necessary to perform a patient feature vector construction operation on the input data (patient characteristic data and drug model data; patient characteristic data includes age, gender, BMI, disease course, surgical history, treatment history, historical medication, daily dietary habits, etc.; drug model data includes drug type, name, dosage, and duration of use; food model data includes daily dietary habits, food quantity, and frequency of meals, etc.). The patient characteristic data is then concatenated with the drug model data and food model data respectively to obtain the corresponding drug recommendation-related features and food component decomposition results (all with feature labels to facilitate connection with the model). The patient food and drug recommendation module is used to obtain a list of recommended medications and specific food patterns for patients, and then push the list of recommended medications and specific food patterns to relevant personnel for processing.

[0025] In this embodiment, the specific process of pushing the patient's recommended medication list and specific food patterns to relevant personnel for reference is as follows: The patient's recommended medication list and specific food patterns are merged to generate a patient condition report, which is then pushed to relevant personnel for processing. The patient condition report includes the recommended medication list, dietary recommendations, and precautions. It should be noted that the recommended medication list is ordered by priority and requires the inclusion of evidence and risk information for each medication. The dietary recommendations include foods to be avoided and recommended food patterns. The precautions include drug-food interactions and monitoring recommendations.

[0026] Based on the above implementation details, this embodiment has the following advantages over the prior art: This embodiment acquires state data of various foods and drugs after they pass through the digestive tract, patient endoscopic image data, patient characteristic data, and mucosal damage data; it performs a healing impact analysis of various foods and drugs based on the state data of various foods and drugs after they pass through the digestive tract; it acquires the healing impact analysis results of various foods and drugs, vascular image contrast data in patient endoscopic images, the ratio of vascular pixel count to total mucosal pixel count, vascular branch count data, and damage area and depth data of each damaged area of ​​the mucosa; it performs predictive analysis of the healing status of patient lesions corresponding to various foods and drugs based on the healing impact analysis results of various foods and drugs, the clarity of vascular texture in patient endoscopic images, and mucosal damage data; it establishes a patient food and drug matching analysis model, and recommends patient foods and drugs based on the patient food and drug matching analysis model, the patient healing status prediction analysis results corresponding to various foods and drugs, and patient characteristic data; it acquires a patient drug recommendation list and specific food patterns. The system pushes recommended medication lists and specific food patterns to relevant personnel for processing; it analyzes the state data of food and medications after they pass through the digestive tract, quantifying the potential mechanical stimulation or burden on fragile and damaged colonic mucosa by calculating sharpness and residue load; it provides a more sensitive and accurate quantitative tool by automatically extracting microscopic features such as vascular contrast, vascular pixel ratio, and number of vascular branches from endoscopic images; it not only predicts drug efficacy but also predicts mucosal healing trends under specific food patterns; regarding medication recommendations, it integrates machine learning models to predict the probability of remission and duration of remission, and combines these with the aforementioned healing predictions for weighted integration, generating a medication recommendation list sorted by a comprehensive evaluation index, while providing evidence and risk explanations for the recommendation results, enhancing the transparency and safety of clinical decision-making; regarding dietary recommendations, it combines association rule mining and Bayesian networks to obtain specific food pattern suggestions, which are tailored to the patient's individual history and predicted healing trajectory, rather than a universal diet, greatly improving the pertinence and patient suitability of dietary management.

[0027] The specific steps for implementing the corresponding functions of each unit module in the information-based planning and evaluation system for the efficacy of active ulcerative colitis described above can be found in the embodiments of the information-based planning and evaluation system for the efficacy of active ulcerative colitis described above, and will not be repeated here.

[0028] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. An information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis, characterized in that, include: The colitis data acquisition module is used to acquire data on the state of various foods and drugs after they pass through the digestive tract, patient endoscopic image data, mucosal damage data, and patient characteristic data. The Recovery Impact Analysis module is used to analyze the recovery impact of various foods and drugs based on the state data of various foods and drugs after passing through the digestive tract. The healing prediction and analysis module is used to predict and analyze the healing status of patients' lesions based on the results of the analysis of the impact of various foods and drugs on healing, patient endoscopic image data, and mucosal damage data. The patient food and drug matching analysis module is used to establish a patient food and drug matching analysis model, and to make recommendations on patient food and drugs based on the patient food and drug matching analysis model, the prediction analysis results of patient healing status corresponding to various foods and drugs, and patient characteristic data. The patient food and drug recommendation module is used to obtain a list of recommended medications and specific food patterns for patients, and then push the list of recommended medications and specific food patterns to relevant personnel for processing.

2. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 1, characterized in that, The analysis process for the impact of various foods and medicines on recovery is as follows: The sharpness analysis results of various foods and drugs were obtained by using data on the hardness decay of various foods and drugs after passing through the digestive tract and shape contour image data. The analysis results of the residue load of various foods and drugs were obtained by using data on the content of insoluble dietary fiber and undigested particulate matter after various foods and drugs passed through the digestive tract. Based on the analysis results of the sharpness and residue load of various foods and medicines, the healing impact of various foods and medicines was analyzed.

3. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 2, characterized in that, The process of obtaining the sharpness analysis results of various foods and drugs from the data on the hardness attenuation and shape contour images after passing through the digestive tract includes the following specific steps: Analyzing the sharpness of various foods and drugs based on the data on the hardness attenuation and shape contour images after passing through the digestive tract, wherein the sharpness analysis formula for the i-th type of food and drug is: Let i be the number corresponding to each food and medicine, where i can be any number from 1 to N, and N be the maximum number corresponding to each food and medicine. Let i be the rate at which the hardness of the i-th type of food and medicine decreases. Let be the shape parameters of the hardness decay curve for the i-th type of food and medicine. For reference to the softening half-life, Let be the minimum radius of curvature corresponding to the shape contour line of the i-th type of food and medicine. and These are the weights corresponding to the hardness decay and the minimum radius of curvature, respectively.

4. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 3, characterized in that, The analysis results of the residue load of various foods and drugs obtained from the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after various foods and drugs have passed through the digestive tract include the following specific steps: The residue load of various foods and drugs is analyzed based on the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after various foods and drugs have passed through the digestive tract. The analysis process is as follows: the data on the content of insoluble dietary fiber and the content of indigestible particulate matter after various foods and drugs have passed through the digestive tract are divided by the reference content of insoluble dietary fiber and the content of indigestible particulate matter, respectively, to obtain the food residue load and the drug residue load. The food residue load and the drug residue load are weighted and summed to obtain the analysis results of the residue load of various foods and drugs.

5. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 4, characterized in that, The analysis of the healing impact of various foods and medicines includes the following specific steps: The analysis of the healing impact of various foods and medicines is conducted based on the results of the sharpness analysis and residue load analysis. The process of the healing impact analysis of various foods and medicines is as follows: The results of the sharpness analysis and residue load analysis of various foods and medicines are obtained; the results of the sharpness analysis and residue load analysis of various foods and medicines are weighted and summed to obtain the final result of the healing impact analysis of various foods and medicines.

6. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 5, characterized in that, The specific process for predicting and analyzing the healing status of patient lesions corresponding to various foods and medicines is as follows: Obtain the healing impact analysis results of various foods and medicines, vascular image contrast data, the ratio of vascular pixels to total mucosal pixels, vascular branch count data, and the area and depth of each mucosal lesion area based on the vascular image contrast analysis results, vascular texture clarity data, and mucosal lesion data in the patient's endoscopic images. Specifically, the predictive analysis process for the healing status of patient lesions corresponding to various foods and medicines involves: normalizing the vascular image contrast data, the ratio of vascular pixels to total mucosal pixels, and the vascular branch count data in the patient's endoscopic images, and then weighting and summing the processed results. The following steps were taken: First, the clarity of the patient's blood vessels was determined. Then, the area of ​​each damaged mucosa region was summed to obtain the total area of ​​the damaged mucosa. Dividing this total area by the total mucosal area yielded the percentage of mucosal damage. The average depth of each damaged mucosa region was calculated. Dividing this average depth by the maximum depth of each damaged region yielded the total depth of the mucosal damage. Finally, the degree of mucosal damage was determined by summing the depth of damage and the percentage of damage. The reciprocal of the clarity of the blood vessels was added to the degree of mucosal damage to determine the severity of the affected area. Finally, the severity of the affected area was multiplied by the results of the analysis of the impact of various foods and medications on healing to obtain the predictive analysis results for the healing of the affected area corresponding to each food and medication.

7. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 6, characterized in that, The specific process of establishing a patient food and drug matching analysis model, and recommending patient food and drugs based on the model, the prediction analysis results of patient healing status corresponding to various foods and drugs, and patient characteristic data, is as follows: The patient food and drug matching analysis model includes a patient drug recommendation sub-model and a food suggestion sub-model. The drug recommendation sub-model has four layers: a patient drug model data preprocessing layer, a drug recommendation basic prediction model layer, a basic prediction result integration layer, and an explanation and risk warning layer. The food suggestion sub-model has four layers: a patient food model data preprocessing layer, an association rule mining and Bayesian network layer, and a diet... The system consists of a pattern generation layer and a diet plan construction layer. The drug recommendation sub-model construction process involves: processing patient feature data and drug model data in the patient-drug model data preprocessing layer to transform them into drug recommendation-related features; standardizing these features; constructing and training two basic random forest models in the basic prediction result integration layer; inputting the standardized drug recommendation-related features into the two basic random forest models to predict the duration of remission and the probability of remission for various drugs; and comparing the predicted duration of remission, the probability of remission, and the patient's healing status. The predictive analysis results are weighted and averaged in the basic prediction result integration layer to obtain a comprehensive evaluation index for various drugs. Simultaneously, drugs are ranked according to their comprehensive evaluation indices to obtain a recommended drug list for each patient. In the risk warning explanation layer, explanations and risk warnings are added to the recommended drugs in the patient's drug list. The food suggestion sub-model construction process is as follows: patient feature data and food model data are processed for patient dietary structure in the patient food model data preprocessing layer to obtain food component decomposition results. In the association rule mining and Bayesian network layer, association rule mining is used to obtain various food component decomposition results and their correlation with patient healing. The correlation between the prediction and analysis results of the situation is analyzed, and a Bayesian network is constructed to establish the conditional dependency relationship between food components and the prediction and analysis results of patient healing. Based on the correlation between the decomposition results of various food components and the prediction and analysis results of patient healing, as well as the conditional dependency relationship between food components and the prediction and analysis results of patient healing, food component suggestions are generated for patients at the diet pattern generation layer. Based on the patient's food component suggestions, specific food patterns are generated at the diet plan construction layer. The prediction and analysis results of patient healing and patient characteristic data corresponding to various foods and drugs are input into the patient food and drug matching analysis model to obtain the patient drug recommendation list and specific food patterns.

8. The information-based planning and evaluation system for the treatment efficacy of active ulcerative colitis as described in claim 7, characterized in that, The specific process of pushing the patient's medication recommendation list and specific food patterns to relevant personnel for reference is as follows: merging the patient's medication recommendation list and specific food patterns to generate a patient condition report, and pushing the patient condition report to relevant personnel for processing; the patient condition report includes the medication recommendation list, dietary suggestions, and precautions.

Images