Intelligent analysis method and system for intervention effect of colitis based on multi-modal biological data

Abstract

This invention discloses an intelligent analysis method and system for the intervention effect of colitis based on multimodal biological data. By associating multimodal biological data with spatial regions of colonic tissue, regionalized biological state data is established. Combined with the intervention time sequence, the evolution trajectory of inflammatory state in each region is constructed, enabling independent analysis of the inflammatory changes in different regions and avoiding the masking of local abnormalities by overall indicator evaluation. By analyzing the relationship between inflammatory state changes in different regions, local inflammatory migration areas and abnormal recovery areas are identified, and persistent damage or newly occurring inflammatory areas that still exist under the overall improvement trend are discovered, avoiding misjudgment of treatment effect. Furthermore, the overall evaluation results are corrected by utilizing local abnormal areas, so that the final result simultaneously reflects the overall degree of recovery and local spatial evolution characteristics, improving the accuracy and reliability of colitis intervention effect evaluation.

Description

Technical Field

[0001] This invention relates to the field of intelligent analysis, and more specifically to an intelligent analysis method and system for the intervention effect of colitis based on multimodal biological data. Background Technology

[0002] With the development of biodetection and omics analysis technologies, multimodal biological data-based methods for evaluating the effectiveness of colitis interventions are increasingly being applied in drug screening, treatment optimization, and disease mechanism research. These methods typically begin by collecting various types of biological data before and after colitis intervention, including clinical data reflecting phenotypic changes such as weight and disease activity values, colonic pathological images reflecting the degree of tissue damage, cytokine expression data reflecting the immune inflammatory state, and gut microbiota composition, metabolomics, and gene expression data reflecting changes in the gut microbiota. Subsequently, the biological data from different sources and at different scales are standardized, and through feature extraction and data fusion, the changes at different biological levels are converted into comprehensive disease state characteristics to characterize the impact of interventions on inflammation levels, mucosal repair, microbiota recovery, and metabolic function improvement. Furthermore, a comprehensive evaluation index is constructed based on the changes in multimodal characteristics before and after intervention to determine the effectiveness of the treatment plan, thus achieving a multidimensional evaluation of the colitis intervention effect. By integrating information from multiple biological levels, this method, compared to evaluation methods relying solely on a single inflammatory marker or pathological score, can more comprehensively reflect the changes in the colitis disease state, improving the accuracy and reliability of intervention effect evaluation.

[0003] However, in actual colitis intervention, the disease state is not uniformly distributed throughout the colon. Instead, it may manifest as local remission, persistent inflammation, or even inflammatory migration across different intestinal segments during treatment. For example, some areas may show a significant reduction in inflammation levels after intervention, while others may remain highly inflammatory due to differences in drug action, local microbiota imbalance, or changes in the immune microenvironment, or even the emergence of new inflammatory clusters. Existing evaluation methods based on multimodal biological data typically involve fusion analysis after obtaining overall pathological scores, average inflammatory indicators, or global omics characteristics. This compresses the biological response information from different regions into a single comprehensive evaluation result, causing abnormal changes in local areas to be masked by the overall improvement trend. This makes it difficult for the system to identify whether inflammation has spatially migrated or whether high-risk local areas remain in a state of persistent damage. Therefore, in special cases such as asynchronous recovery of inflammatory areas, local inflammatory metastasis, and enhanced spatial heterogeneity, existing methods are prone to misinterpreting overall indicator improvements as overall disease recovery, failing to accurately reflect the true spatial evolution of colitis after intervention. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, this invention provides an intelligent analysis method and system for the intervention effect of colitis based on multimodal biological data, in order to solve the problems existing in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A smart analysis method for the intervention effect of colitis based on multimodal biological data, the method comprising: S1: Acquire multimodal biological data at different time stages during the intervention of colitis, and perform regional correspondence processing on the multimodal biological data according to the spatial distribution of colon tissue to establish regionalized biological status data corresponding to each colon region; S2: Based on regionalized biological state data, correlation analysis was performed on the changes in biological state of each colonic region according to the intervention time sequence to construct the inflammatory state evolution trajectory corresponding to each colonic region. S3: Based on the evolution trajectory of inflammatory state, analyze the relationship between changes in inflammatory state between different colonic regions, and identify local inflammatory migration areas and abnormal recovery areas during the intervention process; S4: Based on the local inflammatory migration area and abnormal recovery area, the overall intervention effect evaluation results are corrected to generate a colitis intervention effect evaluation result that takes into account the spatial evolution characteristics of inflammation.

[0006] This invention also provides an intelligent analysis system for the intervention effect of colitis based on multimodal biological data, including: Data module: Acquire multimodal biological data at different time stages during colitis intervention, and perform regional correspondence processing on the multimodal biological data according to the spatial distribution of colonic tissue to establish regionalized biological status data corresponding to each colonic region; Evolutionary trajectory module: Based on regional biological state data, the changes in biological state of each colon region are correlated according to the intervention time sequence to construct the evolutionary trajectory of inflammatory state corresponding to each colon region; Analysis module: Based on the evolution trajectory of inflammatory state, it analyzes the relationship between changes in inflammatory state between different colonic regions and identifies local inflammatory migration areas and abnormal recovery areas during the intervention process; Intervention module: Based on the local inflammatory migration area and abnormal recovery area, the overall intervention effect evaluation results are corrected to generate colitis intervention effect evaluation results that take into account the spatial evolution characteristics of inflammation.

[0007] The technical effects and advantages of this invention are as follows: By correlating multimodal biological data with the spatial distribution of colonic tissue, regionalized biological state data corresponding to each colonic region is established. Furthermore, based on the temporal changes during the intervention process, an inflammatory state evolution trajectory is constructed, enabling the inflammatory changes in different regions to be independently characterized. This avoids the problem of local abnormal changes being averaged out by traditional methods that rely solely on overall pathological scores or global biological indicators for evaluation. Simultaneously, by analyzing the inflammatory state evolution relationship between different colonic regions, local inflammatory migration areas and abnormal recovery areas that occur during the intervention process can be identified. This allows for the discovery of persistent damage areas or newly formed inflammatory areas that persist even when overall indicators improve, avoiding misjudging local deterioration as overall recovery. Furthermore, by using local inflammatory migration areas and abnormal recovery areas to correct the overall intervention effect evaluation results, the final intervention effect evaluation results can simultaneously reflect the overall recovery degree and the local spatial evolution state, improving the accuracy and reliability of colitis intervention effect analysis. Attached Figure Description

[0008] Figure 1 A flowchart illustrating the intelligent analysis method for the intervention effect of colitis based on multimodal biological data provided in this application embodiment. Detailed Implementation

[0009] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent analysis method and system for the intervention effect of colitis based on multimodal biological data involved in the present invention are not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0010] like Figure 1 As shown, this invention provides an intelligent analysis method for the intervention effect of colitis based on multimodal biological data, including the following steps: Step S1: Acquire multimodal biological data at different time stages during the colitis intervention process, and perform regional mapping processing on the multimodal biological data according to the spatial distribution of colonic tissue to establish regionalized biological state data corresponding to each colonic region. The specific process is as follows: First, during the colitis intervention experiment, multimodal biological data was acquired at different stages before and after the intervention according to preset time points. This multimodal biological data included colonic pathological image data reflecting tissue damage, inflammatory factor data reflecting immune inflammation levels, gut microbiota composition data reflecting changes in gut microbiota, and metabolic data reflecting changes in biological function. Data from different time stages were used to describe the dynamic changes in colonic tissue during the intervention process. For example, initial inflammatory status data was acquired before drug intervention, and pathological images, inflammatory indicators, and microbiota change data were acquired on days 3, 7, and 14 after intervention, respectively, thus forming continuous intervention response data. Subsequently, based on the actual spatial structure of the colonic tissue, the acquired multimodal biological data underwent region mapping processing. That is, the overall biological data was spatially divided according to different anatomical regions of the colon, so that biological data from different sources could be mapped to the same colonic region. For example, the colon was divided into proximal colonic region, mid-colonic region, and distal colonic region. For pathological image data, data was mapped according to the image acquisition position. The regions to which data belong are determined. For inflammatory factor data, the data is associated with the corresponding region based on the sample collection location. For microbiota and metabolic data, a regional mapping relationship is established based on the detection results of the corresponding intestinal segment samples. Through the above regional correspondence processing, the multimodal biological data that originally only reflected the overall state of the colon is transformed into regionalized biological state data with spatial location information. For example, the changes in inflammatory factors, the degree of tissue damage, and the changes in microbiota corresponding to the proximal colon region are obtained, while the same type of data corresponding to the distal colon region is obtained. This forms a multidimensional biological state description of different colon regions at different intervention time stages, providing a data foundation for subsequent analysis of the trajectory of inflammatory state changes in each region and identification of local inflammatory migration.

[0011] Step S2: Based on regionalized biological state data, correlation analysis is performed on the changes in biological state of each colonic region according to the intervention time sequence to construct the evolution trajectory of inflammatory state corresponding to each colonic region. The specific process is as follows: Based on the regionalized biological status data corresponding to each colon region obtained in S1, the biological status data of the same colon region at different stages are first sorted according to the intervention time order, so that the inflammatory change process in the same region forms a continuous data sequence. For example, the levels of inflammatory factors, the degree of pathological damage, changes in microbiota, and metabolic status of the proximal colon region before intervention, on day 3, day 7, and day 14 are sequentially arranged to form a state sequence of changes in this region during the treatment process. Subsequently, correlation analysis is performed on the relationship between changes in biological status at different time points within the same colon region to analyze the direction, magnitude, and sequence of changes of various biological indicators during the intervention process. This is to determine the dynamic change pattern of the inflammatory state of this region from the initial occurrence of inflammation, inflammation relief, to tissue recovery. For example, if a region shows a rapid decrease in inflammatory factors and a slow recovery of pathological damage in the early stage of intervention, it indicates that the region first experiences immune inflammation relief and then enters the tissue repair stage. If a region shows a decrease in inflammatory factors followed by a resurgence, accompanied by increased microbiota dysbiosis, it indicates that the region may have recurrent inflammation or impaired recovery. Furthermore, the results of multimodal changes in biological status corresponding to the same region at different time stages are continuously connected to form the inflammatory state evolution trajectory corresponding to the colon region, which is used to describe the state transformation process of this region throughout the entire intervention process. For example, for the distal colon region, if its regionalized biological state data shows a continuous change from "high inflammation → decreased inflammation → tissue repair," then a corresponding normal recovery trajectory is constructed. Conversely, for the proximal colon region, if the change is from "mild inflammation → increased inflammation → persistent damage," then an abnormal evolutionary trajectory is formed. Through this process, each colon region has an independent corresponding inflammatory state evolutionary trajectory, thus avoiding the problem of existing methods relying solely on overall average indicators to judge the degree of disease recovery and failing to reflect asynchronous changes in different regions. This provides a foundation for subsequent identification of local inflammatory migration areas and abnormal recovery areas.

[0012] Step S3: Based on the inflammatory state evolution trajectory, analyze the relationship between inflammatory state changes in different colonic regions to identify local inflammatory migration areas and abnormal recovery areas during the intervention process. The specific process is as follows: Based on the evolution trajectory of inflammatory state, the relationship between changes in inflammatory state among different colonic regions is analyzed. The multimodal inconsistency value and trajectory deviation value of the region are calculated. The multimodal inconsistency value and trajectory deviation value are added together to obtain the outlier value. Based on the outlier value, the local inflammatory migration region and abnormal recovery region are determined.

[0013] In this embodiment, the calculation steps for the multimodal inconsistency value are as follows: The inflammatory state evolution trajectory corresponding to each colonic region is decomposed according to different biological modalities to obtain the state sequence corresponding to each biological modality at each intervention time stage. For any biological modality, the difference in state values ​​between two adjacent time stages is compared sequentially. When the state value of the later time stage is lower than that of the earlier time stage, the change direction corresponding to the biological modality at that time stage is determined to be the recovery direction. When the state value of the later time stage is higher than that of the earlier time stage, the change direction corresponding to the biological modality at that time stage is determined to be the abnormal direction. When the two are equal, it is determined to be the direction of no change, thus obtaining the change direction sequence corresponding to each biological modality.

[0014] For any two biomodalities within the same colonic region, their directions of change are compared at the same time stage. When the two biomodalities change in the same direction, or both show no change, it is determined that there is no contradictory relationship between them at that time stage. When the two biomodalities change in opposite directions, it is determined that there is a contradictory relationship between them at that time stage, and this contradictory relationship is recorded as a complete contradiction. When one biomodality shows no change and the other shows a recovery or abnormal direction, it is determined that there is a partial contradictory relationship between them at that time stage, and this contradictory relationship is recorded as a partial contradiction.

[0015] Within the same time period, the pairwise combinations of all biological modalities in the colon region are traversed and counted. The number of all completely contradictory relationships and the number of partially contradictory relationships are accumulated and divided by the total number of pairwise combinations of biological modalities in the region to obtain the modal contradiction state value corresponding to the time period. Completely contradictory relationships are counted once, and partially contradictory relationships are counted half a time, so that the modal contradiction state value is limited to between zero and one.

[0016] For the modal contradiction state values ​​corresponding to two adjacent time stages, the absolute value of the difference between them is calculated as the degree of modal contradiction change in the corresponding time stage; the degree of modal contradiction change corresponding to all adjacent time stages is statistically analyzed and used together with the modal contradiction state values ​​corresponding to each time stage as the basis for calculation.

[0017] The modal conflict state values ​​and modal conflict change degrees corresponding to each time stage were normalized. The modal conflict state values ​​and modal conflict change degrees of all time stages were summed and averaged to obtain the multimodal conflict value corresponding to the colon region. The multimodal conflict value is a dimensionless value between zero and one. The larger the value, the more inconsistent the change direction of different biological modes in this region during the intervention process.

[0018] It is important to clarify that, for example, inflammatory factors, microbial status, tissue damage, and metabolic indicators are first sequenced as they change over time. Then, the trends of change in adjacent stages are compared time-by-time. For instance, a decrease in inflammatory factors is recorded as a recovery direction, while an increase in microbial dysbiosis is recorded as an abnormal direction. At the same time point, if a region exhibits "decreased inflammation but worsened microbial status" or "tissue repair but abnormal metabolism," then the different modalities form opposite relationships and are recorded as contradictory, counted as either completely contradictory or partially contradictory. Subsequently, the proportion of contradictions between all pairs of modalities at that time point is calculated to obtain the contradiction state value for that time stage. Further comparison of the changes in this contradiction state value in adjacent time stages is used to identify whether there is a sudden mismatch in biological states. Finally, the contradiction states and their degree of change in each time stage are normalized as a whole to obtain a multimodal contradiction value between 0 and 1. For example, if a region experiences a continuous decrease in inflammation but repeated fluctuations in microbial status and metabolism during treatment, this value will increase, indicating that although the surface indicators of this region have improved, there is a significant inconsistency in the internal biological processes. This allows for the identification of abnormal recovery areas that are difficult to detect using traditional methods.

[0019] The multimodal conflict score is an indicator used to evaluate whether the trends of different types of biological data within the same colonic region are consistent. It measures the coordination of recovery processes at different biological levels, such as inflammatory factors, tissue damage, gut microbiota status, and metabolic function, during colitis intervention. This score does not judge whether a single indicator has improved, but rather analyzes whether multiple biological modalities are changing in the same direction. For example, in a certain region, the level of inflammatory factors decreases after intervention, indicating that the inflammatory response has been alleviated; however, at the same time, the gut microbiota structure continues to deteriorate, indicating that the gut microbiota has not recovered. In this case, the two biological modalities show inconsistent directions of change, indicating a conflict in the recovery process in this region, and the multimodal conflict score will increase. A larger multimodal conflict score indicates a more significant difference in changes between different biological modalities within the region. This means that while some biological states show improvement, other key biological processes remain abnormal, suggesting that the region may have superficial recovery but incomplete internal functional recovery, easily forming an abnormal recovery area. Conversely, a smaller multimodal conflict score indicates that the trends of different biological modalities are more consistent. For example, a decrease in inflammation level, tissue repair, gradual recovery of gut microbiota, and improvement in metabolic function occur simultaneously, indicating that the region is in a stable recovery state. Therefore, multimodal contradiction values ​​can identify local recovery anomalies that still exist even when the overall indicators improve, thus avoiding bias in the judgment of intervention effects caused by relying solely on a single indicator or overall evaluation results.

[0020] In this embodiment, the calculation steps for the trajectory deviation value are as follows: The evolution trajectory of inflammatory state corresponding to each colonic region was divided according to the intervention time sequence, and the change in inflammatory state between adjacent time stages was obtained. Correlation analysis was performed on the change in inflammatory state corresponding to three consecutive time stages. By comparing whether the change direction between two adjacent change quantities is consistent, the continuity of the trajectory in the corresponding time stage was determined. When the change direction of inflammatory state in two adjacent time stages is consistent, the trajectory change in that time period is determined to be continuous. When the change direction of inflammatory state in two adjacent time stages is reversed, the trajectory change in that time period is determined to be deviated. The trajectory direction mismatch value of the corresponding time stage is calculated based on the degree of difference between adjacent change quantities.

[0021] Specifically, the difference between the change in inflammatory state at the current time stage and the change in inflammatory state at the next time stage is calculated to obtain the degree of deviation between the two. The absolute value of the difference between the two changes is used as the numerator, and the sum of the absolute values ​​of the two changes is used as the denominator. The trajectory direction mismatch value corresponding to the current time stage is obtained by the ratio of the absolute value of the difference to the sum of the absolute values. Among them, when the change trends of adjacent time stages are consistent and the change amplitudes are similar, the trajectory direction mismatch value is small, indicating that the inflammatory state in this region is evolving according to a stable trend. When the change trends of adjacent time stages are reversed or the change amplitudes are significantly different, the trajectory direction mismatch value increases, indicating that the evolution process of the inflammatory state in this region deviates from the stable recovery path.

[0022] Based on the trajectory direction mismatch values ​​corresponding to each time stage, the degree of trajectory abnormal enhancement between adjacent time stages is calculated. Specifically, the absolute value of the difference between the trajectory direction mismatch values ​​corresponding to two adjacent time stages is calculated, and the absolute value of the difference is divided by the sum of the trajectory direction mismatch values ​​corresponding to the two time stages to obtain the degree of trajectory abnormal enhancement for the corresponding time stage. The degree of trajectory abnormal enhancement is used to characterize the rate of change of the degree of deviation during the evolution of the inflammatory state. When the trajectory direction mismatch value of adjacent time stages increases rapidly, it indicates that the inflammatory recovery process in this region is gradually deviating from the normal evolution path.

[0023] Based on the trajectory direction mismatch values ​​and the degree of trajectory abnormal enhancement corresponding to all time stages, the trajectory deviation value corresponding to each colon region is calculated. Specifically, the trajectory direction mismatch values ​​corresponding to all time stages are accumulated and divided by the number of corresponding time stages to obtain the overall trajectory direction deviation. The degree of trajectory abnormal enhancement corresponding to all adjacent time stages is accumulated and divided by the number of adjacent time stages to obtain the degree of trajectory deviation change. The overall trajectory direction deviation and the degree of trajectory deviation change are averaged to obtain the trajectory deviation value.

[0024] The trajectory deviation value is a dimensionless value between zero and one, used to characterize the degree of deviation of the inflammatory state evolution trajectory of the colon region from the stable recovery trajectory. The larger the trajectory deviation value, the more obvious the repeated changes in inflammatory state, abnormal recovery rhythm, and trajectory direction reversal phenomena occur in the region during the intervention process, indicating that the region has not evolved according to the normal recovery process and there is a risk of abnormal recovery or local inflammatory migration. The smaller the trajectory deviation value, the more continuous and stable the inflammatory state change process in the region is, which is in line with the expected recovery trend.

[0025] In this embodiment, the steps for determining the local inflammation migration area and abnormal recovery area based on outliers are as follows: The abnormal values ​​corresponding to each colon region are compared with the preset abnormal threshold, and the colon regions with abnormal values ​​not less than the preset abnormal threshold are identified as abnormal candidate regions.

[0026] Obtain the inflammatory state evolution trajectory corresponding to the abnormal candidate region and compare the inflammatory state evolution trajectory with the inflammatory state evolution trajectory corresponding to the adjacent colon region; when the direction of inflammatory state change in the abnormal candidate region is opposite to the direction of inflammatory state change in the adjacent colon region, and the abnormal state gradually transfers from one region to another region, the abnormal candidate region is identified as a local inflammatory migration region.

[0027] When an abnormal candidate region continuously deviates from the normal recovery trend and there is no relationship of inflammatory state change spreading to adjacent regions, the abnormal candidate region is identified as an abnormal recovery region.

[0028] Specifically, the abnormal values ​​corresponding to each colonic region are compared with preset abnormality thresholds. Colonic regions with abnormal values ​​not less than the preset thresholds are identified as candidate abnormality regions, used to screen key analysis areas exhibiting abnormal changes during intervention. For example, the abnormal values ​​corresponding to the proximal colon, transverse colon, and distal colon are compared. When the abnormal value of the transverse colon region reaches or exceeds the preset threshold, while other regions are below the threshold, the transverse colon region is identified as a candidate abnormality region. Subsequent analysis of inflammatory status changes is only performed on this region, reducing the interference of normally recovered regions on the identification of abnormal regions during the overall analysis. The inflammatory status evolution trajectory corresponding to the candidate abnormality region is obtained, and the inflammatory status evolution trajectory is compared with the inflammatory status evolution trajectory corresponding to the adjacent colonic region over time to analyze the direction of inflammatory changes in the candidate abnormality region and the adjacent region at each intervention time stage. When the inflammatory status of the candidate abnormality region continues to rise, while the inflammatory status of the adjacent colonic region changes from rising to falling, and there is a continuous alternation of inflammatory status changes between the candidate abnormality region and the adjacent region, it is determined that the inflammatory status has shifted from one region to another, and the candidate abnormality region is identified as a local inflammatory migration region. For example, if the distal colon region initially exhibits high inflammation levels that gradually decrease with intervention, while the adjacent transverse colon region shows a gradual increase in inflammation levels, it indicates that the inflammation has not completely subsided but has migrated to the transverse colon region. When the inflammatory evolution trajectory of an abnormal candidate region consistently deviates from the normal recovery trend, but its adjacent regions do not show a corresponding increase in inflammation, it is determined that the abnormal candidate region does not have a significant inflammatory migration relationship and is thus identified as an abnormal recovery region. For instance, if the rate of decrease in inflammation level in a certain region is significantly lower than in other regions during intervention, and the states of inflammatory factors and tissue damage remain abnormal for a long period, but the surrounding regions all change according to the recovery trend, it indicates that this region's recovery is hindered, rather than the inflammation spreading to other regions. Through the above methods, abnormal regions are screened based on outliers, and the inflammatory evolution relationship between abnormal regions and adjacent regions is considered to distinguish and identify local inflammatory migration regions and abnormal recovery regions.

[0029] Step S4: Based on the local inflammatory migration area and abnormal recovery area, the overall intervention effect evaluation results are revised to generate a colitis intervention effect evaluation result that considers the spatial evolution characteristics of inflammation. The specific process is as follows: Obtain the overall intervention effect evaluation result based on multimodal biological data, and obtain the local inflammation migration area and abnormal recovery area determined in step S3; count the proportion of local inflammation migration area and abnormal recovery area in all colon areas to obtain the abnormal area ratio; correct the overall intervention effect evaluation result according to the abnormal area ratio. When the abnormal area ratio exceeds the preset ratio, reduce the evaluation level corresponding to the overall intervention effect evaluation result; when the abnormal area ratio does not exceed the preset ratio, keep the overall intervention effect evaluation result unchanged.

[0030] For a identified local inflammation migration area, the evolution trajectory of the corresponding inflammatory state is obtained, and the persistence of the inflammatory state in the local inflammation migration area during the intervention process is calculated. When there is a persistent increase in inflammation in the local inflammation migration area or when the inflammation spreads to adjacent areas, a migration risk correction amount is added to the overall intervention effect evaluation result. For a identified abnormal recovery area, the outlier value of the corresponding area is obtained, and the outlier value of the abnormal recovery area is compared with a preset abnormal recovery threshold. When the outlier value of the abnormal recovery area exceeds the preset abnormal recovery threshold, a recovery anomaly correction amount is added to the overall intervention effect evaluation result.

[0031] The overall intervention effect evaluation results are corrected based on the migration risk correction amount and the recovery abnormality correction amount to obtain the colitis intervention effect evaluation results that take into account the spatial evolution characteristics of inflammation; among them, the colitis intervention effect evaluation results simultaneously characterize the degree of overall inflammation improvement as well as the local inflammation migration and abnormal recovery.

[0032] It should be specifically noted that this step involves a secondary correction of the previously identified local inflammatory migration areas and abnormal recovery areas to avoid masking local abnormalities based solely on evaluation results obtained from overall multimodal data. Specifically, the overall intervention effect evaluation result obtained through multimodal biological data fusion is acquired. This evaluation result reflects the average recovery status across the entire colonic tissue, for example, an evaluation result of "effective intervention" based on the degree of reduction in overall inflammatory factors, the degree of improvement in tissue damage, and the recovery of the microbiota. Simultaneously, the local inflammatory migration areas and abnormal recovery areas identified in S3 are acquired, and the proportion of these abnormal areas to the total number of colonic areas is calculated to obtain the abnormal area proportion, which is used to determine whether there are large-scale local abnormalities in the overall recovery status. For example, if the colon is divided into 10 regions, and 2 regions are identified as local inflammatory migration areas or abnormal recovery areas, then the abnormal area proportion is 20%. When this proportion exceeds the preset proportion, it indicates that although the overall indicators may show improvement, there are still significant spatial abnormalities, thus lowering the original overall intervention effect evaluation level; when the abnormal area proportion is low, it indicates that the abnormal areas only occupy a small area and have a limited impact on the overall recovery trend, thus maintaining the original overall evaluation result. For identified areas of localized inflammation migration, the evolution trajectory of the corresponding inflammatory state is further obtained. Based on the changes in the inflammatory state during different intervention time periods, the degree of persistent abnormality of the inflammatory state in the area is calculated to determine whether the local inflammation is in a state of continuous enhancement. For example, if an area has a high level of inflammation at the beginning of the intervention, then briefly decreases, but continues to rise in subsequent time periods, and adjacent areas also show an increase in the level of inflammation, it indicates that the inflammation in this area is persistent and tends to spread to the surrounding areas. In this case, the corresponding migration risk correction amount is increased. For abnormal recovery areas, the corresponding outliers are obtained and compared with the preset abnormal recovery threshold. When the outlier exceeds the threshold, it indicates that although there is no obvious inflammatory migration in the area, its recovery process deviates significantly from the normal recovery trajectory. For example, the inflammatory indicators decrease but the microbial state remains abnormal, and the tissue repair speed is significantly delayed. In this case, the corresponding abnormal recovery correction amount is increased. Based on the migration risk correction and recovery abnormality correction obtained above, the original overall intervention effect evaluation results are adjusted. When there is local inflammation migration or abnormal recovery, the reliability of the intervention effect evaluation results or the evaluation level is reduced, so that the final evaluation result no longer only reflects the overall average improvement, but also considers the abnormal changes in local areas. For example, if the overall inflammation score shows that the treatment effect is good, but there are obvious inflammation migration and recovery abnormalities in some areas, the corrected evaluation result will not be directly judged as completely effective, but will reflect the intervention effect status with local risks, thereby generating a colitis intervention effect evaluation result that simultaneously includes the degree of overall inflammation improvement, local inflammation migration, and abnormal recovery.

[0033] This invention provides an intelligent analysis system for the intervention effect of colitis based on multimodal biological data, including: Data module: Acquire multimodal biological data at different time stages during colitis intervention, and perform regional correspondence processing on the multimodal biological data according to the spatial distribution of colon tissue to establish regionalized biological status data corresponding to each colon region.

[0034] Evolutionary trajectory module: Based on regional biological state data, the changes in biological state of each colon region are correlated according to the time sequence of intervention, and the evolutionary trajectory of inflammatory state corresponding to each colon region is constructed.

[0035] Analysis module: Based on the evolution trajectory of inflammatory state, it analyzes the relationship between changes in inflammatory state between different colonic regions and identifies local inflammatory migration areas and abnormal recovery areas during the intervention process.

[0036] Intervention module: Based on the local inflammatory migration area and abnormal recovery area, the overall intervention effect evaluation results are corrected to generate colitis intervention effect evaluation results that take into account the spatial evolution characteristics of inflammation.

[0037] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An intelligent analysis method for the intervention effect of colitis based on multimodal biological data, characterized in that, include: S1: Acquire multimodal biological data at different time stages during the intervention of colitis, and perform regional correspondence processing on the multimodal biological data according to the spatial distribution of colon tissue to establish regionalized biological status data corresponding to each colon region; S2: Based on regionalized biological state data, correlation analysis was performed on the changes in biological state of each colonic region according to the intervention time sequence to construct the inflammatory state evolution trajectory corresponding to each colonic region. S3: Based on the evolution trajectory of inflammatory state, analyze the relationship between changes in inflammatory state between different colonic regions, and identify local inflammatory migration areas and abnormal recovery areas during the intervention process; S4: Based on the local inflammatory migration area and abnormal recovery area, the overall intervention effect evaluation results are corrected to generate a colitis intervention effect evaluation result that takes into account the spatial evolution characteristics of inflammation.

2. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 1, characterized in that: The steps for analyzing the relationship of inflammatory status changes among different colonic regions and identifying local inflammatory migration areas and abnormal recovery areas during the intervention process are as follows: Based on the evolution trajectory of inflammatory state, the relationship between changes in inflammatory state among different colonic regions is analyzed. The multimodal inconsistency value and trajectory deviation value of the region are calculated. The multimodal inconsistency value and trajectory deviation value are added together to obtain the outlier value. Based on the outlier value, the local inflammatory migration region and abnormal recovery region are determined.

3. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 2, characterized in that, The steps for calculating the multimodal conflict value are as follows: The inflammatory state evolution trajectory corresponding to each colon region is decomposed according to different biological modalities to obtain the state sequence corresponding to each biological modality at each intervention time stage. For any biological modality, the difference in state values ​​between two adjacent time stages is compared sequentially. When the state value of the later time stage is lower than that of the earlier time stage, the change direction corresponding to the biological modality at that time stage is determined to be the recovery direction. When the state value of the later time stage is higher than that of the earlier time stage, the change direction corresponding to the biological modality at that time stage is determined to be the abnormal direction. When the two are equal, it is determined to be the direction of no change, thus obtaining the change direction sequence corresponding to each biological modality. For any two biomodalities within the same colonic region, their directions of change are compared at the same time stage. When the two biomodalities change in the same direction, or both show no change, it is determined that there is no contradictory relationship between them at that time stage. When the two biomodalities change in opposite directions, it is determined that there is a contradictory relationship between them at that time stage, and this contradictory relationship is recorded as a complete contradiction. When one biomodality shows no change and the other shows a recovery or abnormal direction, it is determined that there is a partial contradictory relationship between them at that time stage, and this contradictory relationship is recorded as a partial contradiction. Within the same time period, the pairwise combinations of all biological modalities in the colon region are traversed and counted. The number of all completely contradictory relationships and the number of partially contradictory relationships are accumulated and divided by the total number of pairwise combinations of biological modalities in the region to obtain the modal contradiction state value corresponding to the time period. Calculate the multimodal conflict value based on the modal conflict state value.

4. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 3, characterized in that, The steps for calculating multimodal conflict values ​​based on modal conflict state values ​​are as follows: For the modal contradiction state values ​​corresponding to two adjacent time stages, the absolute value of the difference between them is calculated as the degree of modal contradiction change in the corresponding time stage; the degree of modal contradiction change corresponding to all adjacent time stages is statistically analyzed and used together with the modal contradiction state values ​​corresponding to each time stage as the basis for calculation. The modal conflict state values ​​and modal conflict change degrees corresponding to each time stage are normalized. The modal conflict state values ​​and modal conflict change degrees of all time stages are summed and averaged to obtain the multimodal conflict value corresponding to the colon region.

5. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 2, characterized in that: The steps for calculating the trajectory deviation are as follows: The evolution trajectory of inflammatory state corresponding to each colon region was divided according to the intervention time sequence, and the amount of change in inflammatory state between adjacent time stages was obtained. Correlation analysis was performed on the amount of change in inflammatory state corresponding to three consecutive time stages. By comparing whether the change direction between two adjacent change quantities is consistent, the continuity of the trajectory in the corresponding time stage was determined. When the change direction of inflammatory state in two adjacent time stages is consistent, the trajectory change in that time period is determined to be continuous. When the change direction of inflammatory state in two adjacent time stages is reversed, the trajectory change in that time period is determined to be deviated. The trajectory direction mismatch value of the corresponding time stage is calculated based on the degree of difference between adjacent change quantities. Based on the trajectory direction mismatch value corresponding to each time stage, the degree of trajectory abnormal enhancement between adjacent time stages is calculated. When the trajectory direction mismatch value increases rapidly, it indicates that the inflammation recovery process in this area is gradually deviating from the normal evolution path. The trajectory direction mismatch values ​​corresponding to all time stages are accumulated and divided by the number of corresponding time stages to obtain the overall deviation of the trajectory direction. The degree of trajectory anomaly enhancement corresponding to all adjacent time stages is accumulated and divided by the number of adjacent time stages to obtain the degree of trajectory deviation change. The trajectory deviation value is obtained by averaging the overall deviation of the trajectory direction with the change in trajectory deviation.

6. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 5, characterized in that: Trajectory direction mismatch value and The difference between the change in inflammation status at the current time stage and the change in inflammation status at the next time stage is calculated to obtain the degree of deviation between the two. The absolute value of the difference between the two changes is used as the numerator, and the sum of the absolute values ​​of the two changes is used as the denominator. The trajectory direction mismatch value at the current time stage is obtained by the ratio of the absolute value of the difference to the sum of the absolute values. The absolute value of the difference between the trajectory direction mismatch values ​​corresponding to two adjacent time stages is calculated, and the absolute value of the difference is divided by the sum of the trajectory direction mismatch values ​​corresponding to the two time stages to obtain the degree of trajectory anomaly enhancement at the corresponding time stage.

7. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 2, characterized in that: The steps for determining local inflammatory migration areas and abnormal recovery areas based on outliers are as follows: The abnormal values ​​corresponding to each colon region are compared with the preset abnormal threshold, and the colon regions with abnormal values ​​not less than the preset abnormal threshold are identified as abnormal candidate regions. Obtain the inflammatory state evolution trajectory corresponding to the abnormal candidate region and compare the inflammatory state evolution trajectory with the inflammatory state evolution trajectory corresponding to the adjacent colon region; when the direction of inflammatory state change in the abnormal candidate region is opposite to the direction of inflammatory state change in the adjacent colon region, and the abnormal state gradually transfers from one region to another region, the abnormal candidate region is identified as a local inflammatory migration region. When an abnormal candidate region continuously deviates from the normal recovery trend and there is no relationship of inflammatory state change spreading to adjacent regions, the abnormal candidate region is identified as an abnormal recovery region.

8. The intelligent analysis method for the intervention effect of colitis based on multimodal biological data according to claim 1, characterized in that: The steps to generate evaluation results for colitis intervention effects that take into account the spatial evolution characteristics of inflammation are as follows: The overall intervention effect evaluation results based on multimodal biological data are obtained, and the local inflammatory migration areas and abnormal recovery areas identified in S3 are obtained; the proportion of local inflammatory migration areas and abnormal recovery areas in all colonic areas is calculated to obtain the abnormal area ratio; the overall intervention effect evaluation results are corrected according to the abnormal area ratio. When the abnormal area ratio exceeds the preset ratio, the evaluation level corresponding to the overall intervention effect evaluation results is reduced; when the abnormal area ratio does not exceed the preset ratio, the overall intervention effect evaluation results remain unchanged. For a given local inflammatory migration area, the evolution trajectory of the inflammatory state in the corresponding area is obtained, and the persistence of the inflammatory state in the local inflammatory migration area during the intervention process is calculated. When there is persistent inflammation or the inflammation spreads to adjacent areas in the local inflammatory migration area, a migration risk correction amount is added to the overall intervention effect evaluation result; for the identified abnormal recovery area, the abnormal value of the corresponding area is obtained, and the abnormal value of the abnormal recovery area is compared with the preset abnormal recovery threshold. When the abnormal value of the abnormal recovery area exceeds the preset abnormal recovery threshold, a recovery abnormality correction amount is added to the overall intervention effect evaluation result.

9. An intelligent analysis system for the intervention effect of colitis based on multimodal biological data, used to implement the intelligent analysis method for the intervention effect of colitis based on multimodal biological data as described in any one of claims 1-8, characterized in that: The system includes: Data module: Acquire multimodal biological data at different time stages during colitis intervention, and perform regional correspondence processing on the multimodal biological data according to the spatial distribution of colonic tissue to establish regionalized biological status data corresponding to each colonic region; Evolutionary trajectory module: Based on regional biological state data, the changes in biological state of each colon region are correlated according to the intervention time sequence to construct the evolutionary trajectory of inflammatory state corresponding to each colon region; Analysis module: Based on the evolution trajectory of inflammatory state, it analyzes the relationship between changes in inflammatory state between different colonic regions and identifies local inflammatory migration areas and abnormal recovery areas during the intervention process; Intervention module: Based on the local inflammatory migration area and abnormal recovery area, the overall intervention effect evaluation results are corrected to generate colitis intervention effect evaluation results that take into account the spatial evolution characteristics of inflammation.

Images