Wound healing state evaluation method and system based on artificial intelligence

Abstract

The invention discloses a wound healing state evaluation method and system based on artificial intelligence, and belongs to the field of wound healing evaluation. The method comprises the following steps: integrating medical record data and wound physiological parameters, and extracting relevance between a medication history and an infection type to form a multi-dimensional pathological factor; hemoglobin oxygenation, blood flow and metabolic heat are subjected to cross-period correlation analysis to construct dynamic physiological time-space information, pathological factors and the physiological time-space information are fused into a heterogeneous graph, nodes comprise drug intensity quantized values, infection risk levels and microcirculation fluctuation values, and edge weights are determined by the time sequence matching degree of drug doses and blood flow changes; based on the node weight and the edge association relationship, the treatment time and the physiological parameter acquisition period are aligned to generate a spatio-temporal evolution matrix, the spatial distribution of the infection risk level and the metabolic heat value characteristics are combined, the dynamic association of the drug action intensity and the microcirculation fluctuation is analyzed, and wound healing evaluation information is generated. The accuracy of wound healing state evaluation can be improved.

Description

Technical Field

[0001] The present application relates to the field of wound healing assessment, and in particular to an artificial intelligence-based wound healing status assessment method and system. Background Art

[0002] Accurate assessment of wound healing status is clinically important for chronic wound management and postoperative recovery, directly impacting treatment optimization and healing prediction. With the advancement of intelligent medical technology, dynamic monitoring and quantitative analysis of wound healing have become crucial for improving diagnosis and treatment efficiency. This is particularly true for complex, infected wounds or multi-stage treatment scenarios, where precise assessment can significantly reduce the risk of complications.

[0003] Existing technologies primarily rely on manual observation combined with imaging to assess wound status. Medical personnel visually observe changes in wound color, measure wound area, and assess tissue firmness through palpation. Some solutions employ multispectral imaging or infrared thermal imaging to indirectly reflect wound oxygenation status by capturing hemoglobin spectral characteristics or differences in tissue temperature distribution.

[0004] Existing wound healing status assessment methods rely on manual experience or isolated detection indicators, and have a single assessment dimension, resulting in a lack of objectivity and continuity in the assessment results. In addition, they are unable to analyze the influencing factors and true healing status of wound healing from the perspective of pathological mechanisms. Therefore, the wound healing status assessment methods in the existing technology have the technical problem of poor accuracy in wound healing status assessment. Summary of the Invention

[0005] This application provides an artificial intelligence-based wound healing status assessment method, system, device and computer storage medium, which can improve the accuracy of wound healing status assessment.

[0006] In a first aspect, the present application provides a method for evaluating wound healing status based on artificial intelligence, the method comprising:

[0007] Obtain the target patient's medical history data and physiological parameter data of the wound area. The medical history data includes treatment time, medication history, and infection type. The physiological parameter data includes hemoglobin oxygenation status values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at multiple time points.

[0008] Based on the correlation between medication history and infection type in medical records, we extract multidimensional pathological factor information related to wound pathological changes. This multidimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division.

[0009] Perform cross-time correlation analysis on the hemoglobin oxygenation status values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at different time points in the physiological parameter data to generate physiological spatiotemporal information;

[0010] Multi-dimensional pathological factor information is integrated with physiological spatiotemporal information to construct a pathophysiological heterogeneous graph. The nodes of the pathophysiological heterogeneous graph include quantified values ​​of drug effect intensity, infection spread risk levels, and microcirculatory blood flow fluctuation values ​​in physiological spatiotemporal information. The edges of the pathophysiological heterogeneous graph represent the correlation between drug effect intensity and microcirculatory blood flow fluctuation values. The weights of the edges are determined based on the temporal matching degree between drug dosage and blood flow changes.

[0011] Based on the weights of the nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are periodically aligned to generate a spatiotemporal evolution matrix;

[0012] Based on the correlation between the quantitative value of drug action intensity and the microcirculatory blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value, wound healing status assessment information is generated.

[0013] In one achievable embodiment, based on the weights of the nodes and edges of the pathophysiological heterogeneous graph, the treatment time is periodically aligned with the collection time of the physiological parameter data to generate a spatiotemporal evolution matrix, including:

[0014] According to the treatment stage division information in the multi-dimensional pathological factor information, the alignment time window length corresponding to the treatment stage is determined. The acute treatment stage uses a first preset time length as the alignment time window length, and the recovery treatment stage uses a second preset time length as the alignment time window length. The first preset time length is less than the second preset time length.

[0015] Based on the weight distribution of nodes and edges in the pathophysiological heterogeneous graph, the treatment time and the acquisition time of physiological parameter data are aligned according to the length of the alignment time window to generate a spatiotemporal evolution matrix.

[0016] In one feasible embodiment, based on the weight distribution of nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are aligned according to the length of the alignment time window to generate a spatiotemporal evolution matrix, including:

[0017] The treatment time is divided into multiple continuous time segments based on the length of the aligned time window. Each time segment contains the quantitative value of the drug effect intensity, infection spread risk level and microcirculatory blood flow fluctuation value of the corresponding stage in the treatment stage division information.

[0018] For the microcirculatory blood flow fluctuation value in each time segment, according to the weight distribution of the edges in the pathophysiological heterogeneous graph, the quantitative value of the drug effect intensity connected by the edges with weight values ​​higher than the preset weight threshold is combined with the maximum blood flow fluctuation value in the corresponding time segment to form the time segment dominant parameter group;

[0019] Sort the drug effect intensity quantification values ​​and infection spread risk level values ​​in the time segment dominant parameter group by their frequency of occurrence within the aligned time window length to generate a parameter priority sequence within each time segment;

[0020] According to the arrangement order of the quantitative values ​​of drug effect intensity, infection spread risk level values, and microcirculatory blood flow fluctuation values ​​in the parameter priority sequence, the collection time points of physiological parameter data are mapped to the time points of treatment time to form a spatiotemporal evolution matrix with time segments as rows and parameter priorities as columns.

[0021] In one achievable embodiment, wound healing status assessment information is generated based on the correlation between the quantitative value of drug effect intensity and the microcirculatory blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value, including:

[0022] According to the spatial distribution characteristics of tissue metabolic heat distribution values ​​in the spatiotemporal evolution matrix, the wound area is divided into the central area, the edge area and the surrounding healthy area.

[0023] According to the correlation between the quantitative value of drug effect intensity and the fluctuation value of microcirculatory blood flow in the spatiotemporal evolution matrix, the correlation strength values ​​of the quantitative value of drug effect intensity and the fluctuation value of microcirculatory blood flow in the central area, the edge area and the surrounding healthy area were calculated respectively;

[0024] Based on the differences in association strength values ​​among the central area, edge area, and surrounding healthy areas, combined with the spatial distribution characteristics of infection spread risk level values, wound healing status assessment information of the wound area is generated.

[0025] In one feasible embodiment, based on the differences in the correlation strength values ​​between the central area, the edge area, and the surrounding healthy area, combined with the spatial distribution characteristics of the infection spread risk level values, wound healing status assessment information of the wound area is generated, including:

[0026] The central region's association strength value is superimposed with the data of the corresponding spatial position in the infection spread risk level value to generate the central region's healing activity parameter. At the same time, the peripheral region's association strength value and the infection spread risk level value are weighted and fused according to the inverse of the time window length in the treatment stage division information to generate the peripheral region's healing inhibition parameter.

[0027] The numerical difference between the healing activity parameter and the healing inhibition parameter within the time window is calculated. When the increase in the healing activity parameter within the continuous time window is greater than the increase in the healing inhibition parameter, the healing status of the central area and the edge area is marked as a positive difference mode.

[0028] The positive difference pattern is matched with the distribution gradient direction of the infection spread risk level value from the central area to the surrounding healthy area. According to the gradient direction of the infection spread risk level value and the growth direction of the healing activity parameter in the positive difference pattern, the wound healing status assessment information of the wound area is determined. The wound healing status assessment information includes the tissue regeneration stage and the inflammation retention stage.

[0029] In one feasible embodiment, based on the correlation between medication history and infection type in the medical record data, multi-dimensional pathological factor information related to wound pathological changes is extracted. The multi-dimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division, including:

[0030] According to the drug type and single drug dosage data in the medication history, combined with the preset correspondence between drug type and benchmark strength, the single drug strength value is calculated to generate drug effect strength information;

[0031] Determine the infection spread risk level value based on the pathogen type corresponding to the infection type and the preset correspondence between the pathogen type and the infection spread level;

[0032] The treatment stage division information is determined according to the frequency of medication in the medication history, and the treatment stage division information includes a recovery treatment stage and an acute treatment stage.

[0033] In one feasible embodiment, multi-dimensional pathological factor information is integrated with physiological spatiotemporal information to construct a pathophysiological heterogeneous graph, including:

[0034] Based on the dosage and frequency data in the drug effect intensity information and the time window length in the treatment stage division information, the drug dosage data is divided into discrete drug effect intensity quantified values ​​according to the time window length, and each time window corresponds to a drug effect intensity quantified value;

[0035] Extract infection type data from the infection spread risk information and, combined with the changing trend of microcirculatory blood flow values ​​within the corresponding time window, map the infection type to different levels of infection spread risk values. The infection spread risk value increases with the decrease in microcirculatory blood flow values.

[0036] The microcirculation blood flow values ​​in the physiological time-space information are arranged in chronological order, and the fluctuation differences between adjacent time points are calculated to obtain the microcirculation blood flow fluctuation value sequence;

[0037] The association between the quantitative value of drug effect intensity and the corresponding time window in the microcirculatory blood flow fluctuation value sequence is established. Each association constitutes an edge of the pathophysiological heterogeneous graph. The weight of the edge is determined by counting the ratio of the direction of drug dose change to the direction of blood flow fluctuation in the same time window. The higher the ratio, the greater the weight of the edge.

[0038] In a second aspect, the present application provides an artificial intelligence-based wound healing status assessment system, the system comprising:

[0039] An acquisition module is used to obtain the target subject's medical history data and physiological parameter data of the wound area. The medical history data includes treatment time, medication history, and infection type. The physiological parameter data includes hemoglobin oxygenation status values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at multiple time points.

[0040] An extraction module is used to extract multidimensional pathological factor information related to wound pathological changes based on the correlation between medication history and infection type in medical records. The multidimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division;

[0041] A generation module is used to perform cross-time correlation analysis on the hemoglobin oxygenation state values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at different time points in the physiological parameter data to generate physiological spatiotemporal information;

[0042] A construction module is used to fuse multi-dimensional pathological factor information with physiological spatiotemporal information to construct a pathophysiological heterogeneous graph. The nodes of the pathophysiological heterogeneous graph include quantified values ​​of drug effect intensity, infection spread risk level values, and microcirculatory blood flow fluctuation values ​​in physiological spatiotemporal information. The edges of the pathophysiological heterogeneous graph represent the correlation between drug effect intensity and microcirculatory blood flow fluctuation values. The weights of the edges are determined based on the temporal matching degree between drug dosage and blood flow changes.

[0043] The generation module is also used to periodically align the treatment time with the collection time of physiological parameter data based on the weights of the nodes and edges of the pathophysiological heterogeneous graph to generate a spatiotemporal evolution matrix;

[0044] The generation module is also used to generate wound healing status assessment information based on the correlation between the quantitative value of the drug effect intensity and the microcirculation blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value.

[0045] In a third aspect, the present application provides an electronic device comprising: a processor, and a memory storing computer program instructions; the processor reads and executes the computer program instructions to implement an artificial intelligence-based wound healing status assessment method as in any embodiment of the first aspect.

[0046] In a fourth aspect, the present application provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, an artificial intelligence-based wound healing status assessment method as in any one of the embodiments of the first aspect is implemented.

[0047] The artificial intelligence-based wound healing status assessment method, system, device and computer storage medium of the present application realize the dynamic association between drug action and microcirculation status by acquiring medical record data including treatment time and medication history and multi-time point physiological parameter data; by extracting multi-dimensional pathological factor information and constructing a pathophysiological heterogeneous graph, the drug action intensity, infection spread risk and microcirculation fluctuation value are time-series matched and weighted, and the synergistic effect of drug dose changes on blood flow fluctuations is quantified, overcoming the limitations of isolated indicator analysis; further combined with the spatial characteristic differences of tissue metabolic heat distribution in the spatiotemporal evolution matrix, a dynamic assessment model for wound area zoning is established, thereby analyzing the differences in healing stages at the pathological mechanism level, significantly improving the accuracy and clinical applicability of wound healing status assessment.

[0048] Furthermore, by dividing the wound area into a central area, an edge area and a surrounding healthy area, and combining the calculation of the correlation intensity between the drug effect intensity and the microcirculation blood flow in the spatiotemporal evolution matrix, as well as the spatial distribution characteristics of the infection spread risk, a multi-region, multi-parameter dynamic correlation analysis was achieved, effectively solving the problem of poor accuracy caused by the single evaluation dimension and lack of continuity of existing technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0050] Figure 1 This is a flow chart of a method for evaluating wound healing status based on artificial intelligence provided by one embodiment of the present application;

[0051] Figure 2 This is a flow chart of a method for generating a spatiotemporal evolution matrix according to an embodiment of the present application;

[0052] Figure 3 This is a flow chart of a method for constructing a pathophysiological isomerization graph provided in one embodiment of the present application;

[0053] Figure 4 This is a schematic diagram of the structure of an artificial intelligence-based wound healing status assessment system provided in one embodiment of the present application;

[0054] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0055] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present application by illustrating the examples of the present application.

[0056] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, the elements defined by the phrase "comprising..." do not exclude the presence of other identical elements in the process, method, article, or device comprising the elements.

[0057] The existing technology mainly uses manual observation combined with imaging examination to evaluate the wound status. Medical staff observe the color change of the wound with the naked eye, measure the wound area, and judge the tissue hardness by palpation. Some technical solutions use multi-spectral imaging or infrared thermal imaging equipment to indirectly reflect the oxygen supply status of the wound by capturing the spectral characteristics of hemoglobin or the difference in tissue temperature distribution. The existing wound healing status assessment method relies on manual experience or isolated detection indicators, and the assessment dimension is single, resulting in a lack of objectivity and continuity in the assessment results. It is also impossible to analyze the influencing factors of wound healing and the actual healing status from the level of pathological mechanisms. Therefore, the wound healing status assessment method in the existing technology has a technical problem of poor accuracy in wound healing status assessment.

[0058] To solve the problems of the prior art, the present invention provides an artificial intelligence-based wound healing status assessment method, system, device, and computer storage medium. The following first introduces the artificial intelligence-based wound healing status assessment method provided in the present invention.

[0059] Figure 1 FIG1 shows a flow chart of a wound healing status assessment method based on artificial intelligence provided by an embodiment of the present application. Figure 1 As shown, the method includes steps S110 to S160.

[0060] S110: Obtain the target subject's medical history data and physiological parameter data of the wound area. The medical history data includes treatment time, medication history, and infection type. The physiological parameter data includes hemoglobin oxygenation status values, microcirculation blood flow values, and tissue metabolic heat distribution values ​​at multiple time points.

[0061] Treatment time refers to the time span from the start of wound treatment to the current assessment time point, including the specific timestamps of each treatment, which is used to mark the timing information of intervention measures such as medication and debridement. Medication history refers to the record of the type of drugs (such as antibiotics, growth factors), single medication dose, frequency and duration used by the target subject during wound treatment, reflecting the intensity of drug intervention on the pathological process of the wound. Infection type refers to the category of wound infection determined based on pathogen detection results (such as bacteria, fungi) or clinical manifestations (such as purulent secretions, range of redness and swelling), which is used to assess the risk of infection spread. The hemoglobin oxygenation status value refers to the quantitative detection of the concentration ratio of oxygenated hemoglobin to deoxygenated hemoglobin in the wound tissue by near-infrared spectroscopy technology, reflecting the oxygen supply efficiency of the local tissue. The microcirculation blood flow value refers to the measurement of blood flow velocity and perfusion volume of capillaries in the wound area based on laser Doppler technology, which characterizes the blood transport capacity of the microvascular network. The tissue metabolic heat distribution value refers to the capture of the surface temperature distribution of the wound by high-resolution infrared thermal imaging equipment, combining the thermodynamic characteristics of tissue metabolic activity and inflammatory response, to quantify the metabolic heat gradient.

[0062] Medical records were extracted through an electronic health record system, including treatment timelines, medication use records, and etiology testing reports. Physiological parameter data were collected using multimodal sensing equipment. Hemoglobin oxygenation status values ​​were measured by a near-infrared spectrometer penetrating the wound surface tissue, and the percentage of oxygenated hemoglobin was calculated through absorption spectral analysis. Microcirculatory blood flow values ​​were obtained by scanning the wound area with a laser Doppler flowmeter, and blood flow velocity and perfusion volume were inverted based on dynamic light scattering signals. Tissue metabolic heat distribution values ​​were obtained by a non-contact infrared thermal imager capturing the wound temperature field and combining it with an ambient temperature compensation algorithm to generate a standardized thermal distribution map. These data were aligned by timestamp to form a time series dataset containing multiple time points and parameters for subsequent correlation analysis.

[0063] For example, burn patients infected with Staphylococcus aureus are targeted, and their medical records include debridement surgery time, intravenous cephalosporin antibiotic records, and wound bacterial culture results. When collecting physiological parameters, a near-infrared spectrometer is used to monitor the oxygenated hemoglobin concentration in the central area of ​​the burn wound daily, with a wavelength range of 650-950nm. A laser Doppler is used to measure the blood flow at the wound edge and surrounding healthy tissues at a grid spacing of 3mm. An infrared thermal imager records the wound temperature distribution every 30 minutes with a resolution of 0.03°C. After denoising and normalization, the data is formed to form a numerical sequence of hemoglobin oxygenation status from the first to the tenth day of treatment, a microcirculation blood flow time series matrix, and a metabolic heat distribution map, which serve as input for the subsequent construction of pathophysiological heterogeneous maps.

[0064] S120: Based on the correlation between medication history and infection type in the medical record data, extract multidimensional pathological factor information related to wound pathological changes. The multidimensional pathological factor information includes drug effect intensity information, infection spread risk information, and treatment stage division information.

[0065] The correlation between medication history and infection type refers to the matching relationship between drug type (such as antibiotics, antifungals) and infection type (such as bacteria, fungi), which is used to evaluate the inhibitory effect of drugs on specific pathogens and potential synergistic or antagonistic effects. Multidimensional pathological factor information refers to a set of multi-level indicators extracted from medical record data that reflects the pathological mechanism of the wound, covering the intensity of drug intervention, dynamic evolution of infection and classification of treatment stages. Drug action intensity information refers to the parameters that quantify the effect of drugs on wound healing, and comprehensively considers the drug type, dose, frequency and mechanism of action to characterize the contribution of the drug to efficacy per unit time. Infection spread risk information refers to the probability of infection spread based on the infection type and pathogen invasiveness assessment, combined with the microcirculation status to predict the possibility of infection spreading to surrounding tissues. Treatment stage division information refers to the treatment cycle stages divided according to the frequency of medication and the trend of changes in wound physiological parameters, such as the acute phase and the recovery phase.

[0066] First, based on the matching relationship between the drug type and the infection type in the medication history (such as antibiotics for bacterial infections), the preset drug and pathogen action weight table is used to calculate the drug's targeted strength for the current infection type and generate drug action strength information. Secondly, according to the pathogen invasiveness level corresponding to the infection type, such as the high invasiveness of Staphylococcus aureus, combined with the preset infection spread risk model, the infection spread risk level value is output. Finally, by analyzing the temporal distribution of medication frequency, such as high-frequency medication in the acute phase and low-frequency medication in the recovery phase, combined with the stage characteristics of the physiological parameters of the wound, such as large fluctuations in blood flow in the acute phase, the treatment process is divided into different stages. After the above information is integrated, multi-dimensional pathological factor information is formed for subsequent heterogeneous graph construction.

[0067] For example, burn patients infected with Staphylococcus aureus are taken as the target objects, and their medication history includes daily intravenous injection of cephalosporin antibiotics. According to the preset association weights between the antibiotic type and bacterial infection, such as the cephalosporin weight is 0.9, the single-day drug effect intensity value is calculated as 0.9 multiplied by the dose, such as 2g. The pathogen invasiveness level corresponding to the infection type is high risk, which is mapped to the infection spread risk level value of level 3. By analyzing the frequency of medication, such as once a day and the characteristics of drastic fluctuations in wound blood flow, it is determined that the current treatment stage is acute. The multi-dimensional pathological factor information finally output includes the drug effect intensity sequence (0.9×2g per day), infection spread risk level 3, and the treatment stage is acute.

[0068] S130: performing a cross-time correlation analysis on the hemoglobin oxygenation state values, microcirculation blood flow values, and tissue metabolic heat distribution values ​​at different time points in the physiological parameter data to generate physiological spatiotemporal information.

[0069] Physiological spatiotemporal information refers to the comprehensive analysis of wound physiological parameters by integrating the dynamic changes of time series with spatial distribution characteristics. This includes the temporal evolution of hemoglobin oxygenation status, the cross-temporal fluctuation pattern of microcirculatory blood flow, and the spatial gradient characteristics of tissue metabolic heat distribution.

[0070] First, dynamic feature extraction is performed along the time dimension. Using a sliding window method, trend fitting is performed on oxygenation values ​​at multiple time points to extract long-term slopes and short-term fluctuations. Long-term slopes can refer to sustained increases and plateaus in the value trend, while short-term fluctuations can indicate sudden increases in the value trend after medication administration. For example, oxygenation efficiency exhibits a sustained upward trend during the initial treatment phase, while short-term fluctuations and decreases occur during the infection phase. Simultaneously, microcirculatory blood flow fluctuations are modeled, and the difference between blood flow values ​​at adjacent time points is calculated to generate a time fluctuation series. This allows identification of periodic features or abnormal fluctuations. Periodic features can include circadian rhythms, while abnormal fluctuations can include sudden drops caused by infection.

[0071] After that, the distribution features of the spatial dimension are extracted, focusing on the coupling relationship between metabolic heat distribution and blood flow. First, metabolic heat zoning can be performed. Based on the tissue metabolic heat distribution value, a spatial clustering algorithm can be used to divide the wound surface into a high-temperature central area, a medium-temperature edge area, and a low-temperature surrounding healthy area to obtain the metabolic heat zoning results. Among them, the high-temperature central area can be a necrotic or infected core area, and the medium-temperature edge area can be an active inflammatory zone. At the same time, metabolic and blood flow coupling analysis is performed to calculate the Pearson correlation coefficient between the metabolic heat distribution and microcirculatory blood flow in each partition to generate a spatial coupling map. For example, high metabolism and low blood flow in the central area indicate a hypoxic state.

[0072] Finally, spatiotemporal features are fused and reduced in dimension, concatenating the temporal and spatial features into a high-dimensional matrix. Principal component analysis is then used to extract key spatiotemporal features, forming a physiological spatiotemporal information matrix with time points as rows and feature dimensions as columns. Temporal features can include trend slopes and fluctuation amplitudes, while spatial features can include subregional temperature mean and coupling coefficients.

[0073] For example, in burn patients infected with Staphylococcus aureus, hemoglobin oxygenation values ​​from day 1 to 10 of treatment showed an average daily increase of 3% during the first three days, followed by a 1.5% decrease on day 4 due to recurrent infection. Microcirculatory blood flow values ​​also surged by 40% within two hours of daily intravenous cephalosporin administration, then decayed by 5% per hour. Tissue metabolic heat distribution was spatially clustered into a high-temperature core zone, a moderate-temperature peripheral zone, and a low-temperature peripheral healthy zone. Further analysis revealed a negative correlation between metabolic heat and microcirculatory blood flow in the core zone, with a correlation coefficient of -0.7, and a weak positive correlation of 0.2 in the peripheral zone. The resulting physiological spatiotemporal information matrix includes oxygenation trend codes, blood flow spike event counts, and the strength of metabolic-blood flow coupling between each zone at each time point. For example, data for day 2 showed an oxygenation trend slope of +0.03, one spike event, and a central zone coupling strength of -0.7.

[0074] S140: Multi-dimensional pathological factor information is integrated with physiological spatiotemporal information to construct a pathophysiological heterogeneous graph. The nodes of the pathophysiological heterogeneous graph include the quantitative value of drug action intensity, the infection spread risk level value, and the microcirculation blood flow fluctuation value in the physiological spatiotemporal information. The edges of the pathophysiological heterogeneous graph represent the correlation between the drug action intensity and the microcirculation blood flow fluctuation value. The weight of the edge is determined based on the temporal matching degree between the drug dosage and the blood flow change.

[0075] The pathophysiological heterogeneous graph is a graph-structured model that integrates multi-source medical data and is used to represent the complex interactions between drug interventions, infection dynamics, and physiological responses during wound healing. Its nodes consist of three types of entities: drug effect strength quantification, infection spread risk grading, and microcirculatory blood flow fluctuation. The drug effect strength quantification represents the standardized numerical value of drug dosage and efficacy; the infection spread risk grading represents the probability of infection spread based on pathogen type and invasiveness assessment; and the microcirculatory blood flow fluctuation represents the absolute or relative difference in blood flow changes between adjacent time points. Edges represent causal relationships or statistical associations between nodes, such as the positive or negative impact of drug effect strength on blood flow fluctuations. Edge weights are quantified by the degree of consistency between changes in drug dosage and the direction of blood flow fluctuations; higher weights indicate stronger correlations.

[0076] First, nodes are generated, including a quantitative value for drug effect intensity, a value for infection spread risk, and a value for microcirculatory blood flow fluctuation. The quantitative value for drug effect intensity can be obtained by multiplying the dosage by a preset drug type weight coefficient, for example, antibiotics have a higher weight coefficient than anti-inflammatory drugs. The infection spread risk value can be mapped to a preset risk level table based on the pathogen type, such as Staphylococcus aureus corresponding to a high risk level. The microcirculatory blood flow fluctuation value can be generated by calculating the difference between blood flow values ​​at adjacent time points.

[0077] Afterwards, edges are established and their weights are calculated. For each time window, the degree of match between the direction of drug dose change (e.g., dose increase, decrease, or maintenance) and the direction of blood flow fluctuation (e.g., increase, decrease) is analyzed. A dynamic time warping algorithm is used to align the drug dose time series with the blood flow fluctuation series, and the proportion of events with the same direction within the same time window is calculated as the edge weight. For example, if a dose increase and a blood flow increase occur simultaneously five times within a time window, accounting for 80% of the total number of events in the window, the edge weight is set to 0.8.

[0078] Finally, the nodes are combined with weighted edges to form a heterogeneous graph topology. Auxiliary edges can also be established between infection spread risk level nodes and blood flow fluctuation nodes. The weight is determined based on the co-occurrence frequency of increased infection level and decreased blood flow to reflect the inhibitory effect of infection on microcirculation.

[0079] S150: Based on the weights of the nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are periodically aligned to generate a spatiotemporal evolution matrix.

[0080] The spatiotemporal evolution matrix is ​​a multidimensional data matrix that integrates temporal dynamics and spatial distribution characteristics. It is used to characterize the coevolution of drug effects, infection risk, and physiological responses during wound healing. The rows of the matrix represent time segments divided by treatment phase, such as daily in the acute phase and weekly in the recovery phase. The columns represent the priority sequence of parameters for drug effect intensity, infection risk level, and blood flow fluctuation within each time segment. The matrix elements contain the quantitative values ​​of the parameters and the strength of their spatiotemporal correlation.

[0081] First, the alignment window length is determined based on treatment stage information. Short windows, such as hourly or daily windows, are used during the acute phase to capture high-frequency fluctuations, while long windows, such as weekly windows, are used during the recovery phase to extract trend features. The window length is dynamically adjusted based on pre-set rules; for example, the acute phase window length is inversely proportional to medication frequency. Next, based on the edge weight distribution within the pathophysiological heterogeneous graph, node parameters connected by edges with weights above a threshold within each time window are selected, such as drug intensity and blood flow fluctuations. These parameters are sorted in descending order of weight to form a dominant parameter group. For example, if the drug-blood flow edge weight in a window is the highest, the drug intensity value for that window is designated as the dominant parameter. Finally, the dominant parameter groups for each time window are arranged in chronological order as matrix rows, and the parameters are sorted by priority, such as weight, as columns. For cross-window parameter conflicts, such as when the same time point belongs to multiple windows, a sliding overlapping window method is used for interpolation and smoothing to ensure temporal continuity. In the resulting matrix, each element contains a parameter value, its associated weight, and a temporal and spatial label, such as a central region location code.

[0082] S160: Based on the correlation between the quantitative value of drug effect intensity and the microcirculatory blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value, wound healing status assessment information is generated.

[0083] Correlation refers to the use of statistical or machine learning methods to reveal the causal or correlational effects of changes in drug dosage on the dynamic fluctuations of blood flow. For example, increasing the antibiotic dosage may increase microcirculatory blood flow. Spatial distribution characteristics refer to the gradient change pattern of infection spread risk level and metabolic heat value in different areas of the wound. For example, the infection risk decreases from the center of the wound to the periphery. Wound healing status assessment information is a quantitative assessment result generated by the fusion of multi-dimensional parameters, including healing stage classification and region-specific healing activity indicators.

[0084] First, regional division and correlation strength calculation are performed. Based on the metabolic heat distribution in the spatiotemporal evolution matrix, a spatial clustering algorithm is used to divide the wound surface into a central zone, a peripheral zone, and a surrounding healthy zone. The Pearson correlation coefficient is calculated for the time series data of drug effect intensity and blood flow fluctuations within each zone to obtain the correlation strength value for each zone. Next, difference analysis and pattern labeling are performed to compare the differences in correlation strength between different zones. For example, a significantly higher correlation strength in the central zone than in the peripheral zone indicates a more pronounced drug effect in the core zone. Incorporating the spatial gradient of infection risk, such as a trend of outward diffusion from the center, if the direction of the infection risk gradient coincides with the direction of increasing correlation strength, the wound is labeled as healing in a positive direction. Finally, wound healing status assessment information is generated. The difference analysis results are integrated with the infection risk distribution and metabolic heat signatures to classify the healing stage using a decision tree or logistic regression model. For example, if the correlation strength in the central zone continues to increase and the infection risk gradient spreads outward, the wound is classified as undergoing tissue regeneration. If the correlation strength in the peripheral zone decreases and the infection risk converges, the wound is classified as undergoing inflammation retention.

[0085] This embodiment achieves a dynamic association between drug effects and microcirculatory status by acquiring medical record data including treatment time and medication history, as well as physiological parameter data at multiple time points. By extracting multidimensional pathological factor information and constructing a pathophysiological heterogeneous graph, the drug effect intensity, infection spread risk, and microcirculatory fluctuation values ​​are time-series matched and weighted, quantifying the synergistic effect of drug dose changes on blood flow fluctuations, overcoming the limitations of isolated indicator analysis. Furthermore, combining the spatial characteristic differences of tissue metabolic heat distribution in the spatiotemporal evolution matrix, a dynamic assessment model for wound area zoning is established, thereby analyzing the differences in healing stages at the pathological mechanism level, significantly improving the accuracy and clinical applicability of wound healing status assessment.

[0086] In one possible implementation, step S150: based on the weights of the nodes and edges of the pathophysiological heterogeneous graph, periodically aligning the treatment time with the collection time of the physiological parameter data to generate a spatiotemporal evolution matrix includes:

[0087] According to the treatment stage division information in the multidimensional pathological factor information, the alignment time window length corresponding to the treatment stage is determined, wherein the acute treatment stage adopts the first preset time length as the alignment time window length, and the recovery treatment stage adopts the second preset time length as the alignment time window length, and the first preset time length is less than the second preset time length; based on the weight distribution of nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are aligned according to the alignment time window length to generate a spatiotemporal evolution matrix.

[0088] Aligning time window lengths refers to a time segmentation strategy dynamically adjusted according to the treatment phase, used to map discrete treatment time points to physiological parameter collection time points on a unified timeline. The acute treatment phase corresponds to the first preset duration, while the recovery treatment phase corresponds to the second preset duration, with the former being shorter than the latter. For example, during the acute phase, an hourly window might be used to capture high-frequency drug interventions and physiological fluctuations, while during the recovery phase, a day-level window might be used to extract long-term trends.

[0089] First, the alignment window length is determined based on the label in the treatment stage classification information, such as acute or recovery. If the current phase is acute, a shorter first preset duration is set, which can be on the order of hours, such as 5 hours, to capture the transient association between drug effects and microcirculatory fluctuations. If the current phase is recovery, a longer second preset duration is set, which can be on the order of days, such as 1 day, to extract the cumulative effect of drug efficacy. Next, based on the node and edge weight distribution in the pathophysiological heterogeneous graph, the treatment timeline and the physiological parameter acquisition timeline are segmented into consecutive time segments according to the determined alignment window length. Each segment contains the quantified value of the drug effect intensity, the infection spread risk level, and the microcirculatory blood flow fluctuation within the corresponding window. For each time segment, node parameters connected by edges with weights above a threshold, such as drug effect intensity and blood flow fluctuation, are selected and sorted in descending order of weight to form a dominant parameter group. Finally, the dominant parameter groups for each time segment are arranged in matrix rows according to the window order, with parameter priority (highest weight) as the column, to generate a spatiotemporal evolution matrix. For time points across windows, sliding window overlapping method is used for interpolation to ensure time series continuity.

[0090] Figure 2 FIG. 1 shows a flow chart of a method for generating a spatiotemporal evolution matrix provided by an embodiment of the present application. Figure 2 As shown, the method includes steps S210 to S240.

[0091] In one feasible embodiment, based on the weight distribution of nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are aligned according to the length of the alignment time window to generate a spatiotemporal evolution matrix, including:

[0092] S210: Divide the treatment time into multiple continuous time segments based on the length of the aligned time window, each time segment containing a quantitative value of the drug effect intensity, an infection spread risk level value, and a microcirculation blood flow fluctuation value of the corresponding stage in the treatment stage division information.

[0093] Based on the length of the aligned time windows, the treatment timeline is divided into continuous time segments. Each time segment contains the corresponding treatment phase division information, including the quantitative value of the drug effect intensity, the infection spread risk level, and the microcirculation blood flow fluctuation value. For example, the quantitative value of the drug effect intensity can be the standardized value of the antibiotic dose, the infection spread risk level can be the high-risk level 3, and the microcirculation blood flow fluctuation value can be the difference in blood flow between adjacent time points. The time segment division adopts the fixed window method, for example, the acute phase is divided into hourly windows, and the recovery phase is divided into daily windows, ensuring that each window contains complete medication records and physiological parameter collection data.

[0094] S220: For the microcirculatory blood flow fluctuation value in each time segment, according to the weight distribution of the edges in the pathophysiological heterogeneous graph, the quantitative value of the drug effect intensity connected to the edges with weight values ​​higher than the preset weight threshold is combined with the maximum blood flow fluctuation value in the corresponding time segment to form a time segment dominant parameter group.

[0095] Weight distribution refers to the statistical characteristics of directed edge weights in a pathophysiological heterogeneous graph, reflecting the central tendency of the correlation between drug effect intensity and microcirculatory blood flow fluctuations in different time windows. The preset weight threshold is a pre-set screening criterion used to filter out low-correlation edges and retain the node parameters connected by high-weight edges.

[0096] For each time segment of microcirculatory blood flow fluctuation, all edges connecting drug intensity nodes and blood flow fluctuation nodes in the pathophysiological heterogeneous graph are traversed. Edges with weights higher than a preset threshold (e.g., 0.7) are screened, and the quantitative values ​​of the drug intensity connected to them are extracted. The maximum blood flow fluctuation value within the time segment is combined with the corresponding drug intensity value to form a dominant parameter group. For example, if the maximum fluctuation value in a time segment is +40 ml / min and the corresponding drug intensity is 0.9 × 2 g, then the dominant parameter group is (0.9 × 2 g, +40 ml / min).

[0097] S230: Sort the drug action intensity quantification values ​​and infection spread risk level values ​​in the time segment dominant parameter group according to the frequency of occurrence within the aligned time window length to generate a parameter priority sequence within each time segment.

[0098] The parameter priority sequence is a sorted list generated by counting the frequency of parameter occurrence within a statistical time window, which represents the priority of the impact of drug effect intensity, infection spread risk level and blood flow fluctuation value on wound healing status.

[0099] Frequency statistics are performed on the drug intensity values ​​and infection spread risk level values ​​within the time segment dominant parameter group. The number of drug intensity values ​​occurring within the aligned time window is calculated. For example, if an antibiotic is used once an hour, the frequency is 24 times per day. The duration of the infection risk level is calculated as the percentage of the window. For example, if a high risk level persists for 6 hours, it accounts for 25% of the window. Parameters are prioritized in descending order of frequency. For example, drug intensity, with the highest frequency, is assigned priority 1, followed by infection risk, and blood flow fluctuation is the lowest.

[0100] S240: According to the arrangement order of the quantitative values ​​of drug effect intensity, infection spread risk level values, and microcirculation blood flow fluctuation values ​​in the parameter priority sequence, the collection time points of the physiological parameter data are mapped with the time points of the treatment time to form a spatiotemporal evolution matrix with time segments as rows and parameter priorities as columns.

[0101] According to the order of the parameter priority sequence, the physiological parameter collection time points (such as the blood flow measurement time) are mapped to the treatment time points (such as the medication time). The timestamp alignment algorithm is used to fill the parameters in the same time window into the matrix columns in priority order. For example, if the time window is 0:00-6:00 on the first day of the acute phase, and the priority sequence is drug intensity, infection risk, and blood flow fluctuation, then the matrix rows correspond to the window, and the columns are filled with the drug intensity mean, infection risk level, and maximum fluctuation value in sequence, forming the row and column structure of the spatiotemporal evolution matrix. The spatiotemporal evolution matrix is ​​shown in Table 1 below. The rows represent the time windows divided by treatment stage, such as the acute phase and the recovery phase, and the window length is dynamically adjusted, such as the acute phase is in hours and the recovery phase is in days. The columns represent the parameter priority sequence, which is sorted from high to low in frequency as drug intensity, infection risk, and blood flow fluctuation.

[0102] Treatment phase Time window Quantitative value of drug effect intensity Infection spread risk level Microcirculatory blood flow fluctuation value Acute phase Day 1 0:00-1:00 0.9×2g (weight 0.85) Level 3 (weight 0.75) +35ml / min (weight 0.7) Acute phase Day 1 1:00-2:00 0.9×2g (weight 0.82) Level 3 (weight 0.73) +28ml / min (weight 0.65) Recovery period Day 4 (all day) 0.7×1g (weight 0.68) Level 2 (weight 0.62) +15ml / min (weight 0.58) Recovery period Day 5 (all day) 0.7×1g (weight 0.65) Level 1 (weight 0.60) +12ml / min (weight 0.55)

[0103] Table 1

[0104] In one practicable embodiment, step S160: generating wound healing status assessment information based on the correlation between the quantitative value of the drug effect intensity and the microcirculatory blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value, includes:

[0105] Based on the spatial distribution characteristics of tissue metabolic heat distribution values ​​in the spatiotemporal evolution matrix, the wound bed area is divided into a central area, a peripheral area, and a surrounding healthy area. Based on the correlation between the quantified drug effect intensity and microcirculatory blood flow fluctuation values ​​in the spatiotemporal evolution matrix, the correlation strength values ​​are calculated for the central area, peripheral area, and surrounding healthy area. Based on the differences in correlation strength values ​​among the central area, peripheral area, and surrounding healthy area, and combined with the spatial distribution characteristics of infection spread risk levels, wound healing status assessment information for the wound bed area is generated. This wound healing status assessment information can include healing stage classification, region-specific healing activity indicators, and dynamic trend prediction. Healing stage classification can include tissue regeneration, inflammation retention, and necrosis stability. Region-specific healing activity indicators can include a central healing activity parameter reflecting the tissue regeneration capacity of the core area, a peripheral healing inhibition parameter representing the degree of inflammation hindering repair, and a metabolic recovery index for the surrounding healthy area.

[0106] The central area, edge area and surrounding healthy area refer to the wound area divided based on the spatial gradient characteristics of the tissue metabolic heat distribution values. The central area has the highest metabolic heat, the edge area has medium metabolism, and the surrounding healthy area has normal tissue temperature. The spatial distribution characteristics of the tissue metabolic heat distribution values ​​refer to the temperature gradient map generated by combining the wound surface temperature field data captured by infrared thermal imaging with the spatial clustering algorithm. The correlation strength value refers to the Pearson correlation coefficient between the quantitative value of the drug action intensity and the microcirculation blood flow fluctuation value, reflecting the degree of causal relationship between drug intervention and blood flow dynamics. The spatial distribution characteristics of the infection spread risk level value refer to the decreasing or increasing trend of the infection risk caused by the invasiveness of the pathogen in different areas of the wound.

[0107] First, based on the metabolic heat distribution data in the spatiotemporal evolution matrix, a spatial clustering algorithm was used to divide the wound surface into three zones. The central zone: High-temperature areas with metabolic heat values ​​exceeding two standard deviations above the overall mean were selected, corresponding to the core of necrosis or severe infection. The peripheral zone: A transition zone with metabolic heat values ​​between one and two standard deviations above the mean, representing areas of active inflammation. The peripheral healthy zone: Metabolic heat values ​​below the mean are consistent with normal tissue temperature. The zoned results were mapped using the spatial coordinates of the thermal distribution map to form zone boundary markers. Next, the correlation strength was calculated. The quantified drug effect intensity values ​​and blood flow fluctuation values ​​within each zone were time-series aligned. The drug intensity and blood flow fluctuation sequences corresponding to the zone were selected from the spatiotemporal evolution matrix. The Pearson correlation coefficient was used to calculate the linear correlation strength between the two. For example, increased drug intensity in the central zone was positively correlated with increased blood flow, while the correlation was weaker in the peripheral zone. The correlation coefficient was mapped to a range of 0–1 to generate a normalized correlation strength value.

[0108] Finally, compare the standard deviation of the association strength values ​​of each region. For example, the association strength of the central area is 0.8, the edge area is 0.3, and the surrounding area is 0.1, with a difference of 0.7. At the same time, match the infection risk space. If the infection risk level decreases from the central area to the periphery, such as level 3 to level 1, and the difference in association strength is greater than the preset threshold, it is determined to be the tissue regeneration stage; if the infection risk is cohesive, such as the risk level in the central area continues to increase, it is marked as the inflammation retention stage. Combined with the difference and the direction of the infection risk gradient, an evaluation report is generated that includes the healing stage classification (such as acute phase-tissue regeneration) and regional activity indicators.

[0109] In one feasible embodiment, based on the differences in the correlation strength values ​​between the central area, the edge area, and the surrounding healthy area, combined with the spatial distribution characteristics of the infection spread risk level values, wound healing status assessment information of the wound area is generated, including:

[0110] The association strength value of the central area is superimposed with the data of the corresponding spatial position in the infection spread risk level value to generate the healing activity parameter of the central area. At the same time, the association strength value of the edge area and the infection spread risk level value are weightedly fused according to the inverse of the time window length in the treatment stage division information to generate the healing inhibition parameter of the edge area; the numerical difference between the healing activity parameter and the healing inhibition parameter within the time window length is calculated. When the growth of the healing activity parameter within the continuous time window is greater than the growth of the healing inhibition parameter, the healing status of the central area and the edge area is marked as a positive difference mode; the positive difference mode is matched with the distribution gradient direction of the infection spread risk level value from the central area to the surrounding healthy area. According to the gradient direction of the infection spread risk level value and the growth direction of the healing activity parameter in the positive difference mode, the wound healing status assessment information of the wound area is determined. The wound healing status assessment information includes the tissue regeneration stage and the inflammation retention stage.

[0111] The healing activity parameter is a comprehensive indicator reflecting the tissue regeneration capacity of the core area of ​​the wound. It characterizes the blood flow recovery and anti-infection capacity of the core area under drug intervention. For example, when the association strength value is high and the infection risk is low, the healing activity parameter increases, indicating active tissue repair. The healing inhibition parameter is a quantitative indicator that characterizes the degree of healing obstruction in the edge area of ​​the wound. It is generated by weighted fusion of the association strength value of the edge area and the infection spread risk level value. The weight coefficient is inversely proportional to the length of the time window of the treatment phase. The shorter the time window, the higher the weight of the inhibitory effect of infection risk on healing. The positive difference pattern is a dynamic pattern that marks the healing trend. When the growth rate of the healing activity parameter in the central area continuously exceeds that of the healing inhibition parameter in the edge area, it indicates that the repair capacity of the core area is dominant and the wound is generally healing. This pattern is identified by comparing the incremental differences between the two types of parameters within continuous time windows.

[0112] The central zone healing activity parameter is obtained by linearly superimposing the association strength value (correlation coefficient between drug effect and blood flow fluctuation) and the infection risk level value at each spatial location in the central zone. For example, if the association strength at a certain location is 0.8 and the infection risk level is level 3, the superposition formula is: activity parameter = 0.8 + 3 × infection risk weight (preset to 0.1), which results in 1.1. The marginal zone healing inhibition parameter is obtained by weighting the association strength value and infection risk level value of the marginal zone by the inverse of the time window length. If the current treatment stage is the acute phase and the time window length is 5 hours, the weight is 1 / 5 = 0.2. The calculation formula is: inhibition parameter = association strength × 0.2 + infection risk level × 0.8. This weighting method emphasizes the inhibitory effect of infection risk in the acute phase.

[0113] The difference between the healing activity parameter and the inhibition parameter is calculated for each time window to obtain the instantaneous difference. For example, if the activity parameter is 1.1 and the inhibition parameter is 0.5, the difference is 0.6. At the same time, the trend of the difference in the continuous time window is analyzed. If the activity parameter increment, such as the activity parameter increment of window 2 increases by 0.3 compared with window 1, is greater than the inhibition parameter increment, such as the inhibition parameter increment is 0.1, it is marked as a positive difference mode. This judgment is achieved by comparing the slope. The slope of the activity parameter must exceed the preset threshold of the inhibition parameter slope, for example, the threshold can be 1.5 times.

[0114] Finally, the spatial gradient direction of the infection spread risk level is extracted. If the risk level decreases from the central area to the surrounding healthy areas, for example, from level 3 to level 2 and then to level 1, the gradient direction is outward diffusion. At the same time, the activity parameter growth direction in the positive difference pattern is matched. For example, if the activity parameter increases from the inside to the outside in the central area and aligns with the infection risk gradient, it is determined to be the tissue regeneration stage. If the activity parameter growth direction is opposite to the risk gradient, for example, the activity in the central area decreases while the risk in the peripheral area increases, it is marked as the inflammation retention stage.

[0115] In one embodiment, the wound healing stage can also be classified into the tissue regeneration stage or the necrosis stabilization stage based on the matching of the judgment results of the positive difference pattern and the direction of the infection risk gradient, combined with the analysis of the difference in the correlation strength of each region. In the tissue regeneration stage, the activity growth of the central area is dominant and the risk expands outward. In the inflammation retention stage, the inhibition of the edge area is enhanced and the risk is concentrated. In the necrosis stabilization stage, there is no obvious change in the activity and inhibition parameters. At the same time, the regional specific activity index is generated by the time series growth trend of the healing activity parameter of the central area, the fluctuation amplitude of the inhibition parameter of the edge area and the metabolic heat recovery rate of the surrounding healthy area. The dynamic trend prediction is based on the incremental ratio of the activity parameter and the inhibition parameter in the continuous time window and the risk gradient evolution rate. The time series model is used to predict the healing process in the next 3-5 days, and the final output includes a comprehensive evaluation report including stage classification, regional activity quantitative indicators and trend curves.

[0116] This embodiment divides the wound area into a central area, an edge area, and a surrounding healthy area. It combines the calculation of the correlation strength between the drug effect intensity and the microcirculation blood flow in the spatiotemporal evolution matrix, as well as the spatial distribution characteristics of the risk of infection spread, to achieve dynamic correlation analysis of multiple regions and multiple parameters, effectively solving the problem of poor accuracy caused by the single evaluation dimension and lack of continuity of existing technologies.

[0117] In one feasible embodiment, step S120: extracting multidimensional pathological factor information related to wound pathological changes based on the correlation between medication history and infection type in the medical record data. The multidimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division information, including:

[0118] Based on the drug type and single medication dosage data in the medication history, combined with the preset correspondence between the drug type and the benchmark intensity, the single medication intensity value is calculated to generate the drug action intensity information; based on the pathogen type corresponding to the infection type, and the preset correspondence between the pathogen type and the infection spread level, the infection spread risk level value is determined; based on the frequency of medication in the medication history, the treatment stage division information is determined, and the treatment stage division information includes the recovery treatment stage and the acute treatment stage.

[0119] The correspondence between drug type and baseline intensity refers to a pre-established mapping table between drug categories and their efficacy intensity, which is used to quantify the intervention ability of different drugs on wound healing. For example, antibiotics may be assigned a higher baseline intensity value, while anti-inflammatory drugs correspond to a medium intensity value. This correspondence is based on clinical trial data or pharmacodynamic research results, and is determined through expert experience or statistical models. The correspondence between pathogen type and infection spread level refers to a preset level classification table based on the invasiveness, drug resistance and risk of infectious complications of the pathogen. For example, Staphylococcus aureus may be mapped to a high-risk level due to its high invasiveness and drug resistance, while Staphylococcus epidermidis corresponds to a medium-risk level. Medication history refers to the drug use information recorded by the patient during wound treatment, including drug name, single dose, route of administration, frequency and duration of medication, which is used to analyze the dynamic impact of drugs on the pathological process of the wound.

[0120] First, based on the drug type in the medication history, such as antibiotics or growth factors, a pre-set table of drug type and baseline strength mappings is retrieved to obtain the baseline strength value for the drug. The single dose is multiplied by the baseline strength value to obtain the single dose strength value. For example, if the baseline strength value for a cephalosporin antibiotic is a pre-set value and the single dose is an intravenous dose, the product of these two values ​​is the strength of action for that single dose. Second, based on the pathogen type associated with the infection type, such as bacteria or fungi, a pre-set table of pathogen type and infection spread level mappings is retrieved to determine the infection spread risk level. For example, if the pathogen is Staphylococcus aureus, it is mapped to a high risk level. Finally, the frequency of medication use in the medication history is analyzed. If the frequency exceeds a pre-set threshold, such as multiple daily dosing, the patient is classified as undergoing acute treatment. If the frequency decreases and stabilizes, such as every other day dosing, the patient is classified as undergoing recovery treatment. Treatment stage division is integrated with the phase-specific characteristics of wound physiological parameters, such as blood flow fluctuations, to ensure that stage division is synchronized with the pathological progression.

[0121] For example, taking a burn patient as an example, his medication history includes daily intravenous injection of cephalosporin antibiotics, and the infection type is Staphylococcus aureus. According to the preset correspondence, the baseline intensity value of cephalosporin antibiotics is high, the single dose is the standard dose, and the calculated single medication intensity value is the product of the baseline intensity and the dose. The pathogen corresponding to the infection type is mapped to a high-risk level, and an infection spread risk level value is generated. The analysis shows that the medication frequency is once a day, and combined with the blood flow parameters showing drastic fluctuations, it is determined that the current treatment stage is the acute phase. The multi-dimensional pathological factor information finally output includes the drug action intensity sequence, the high-risk infection spread level, and the acute phase label.

[0122] Figure 1 FIG. 1 shows a flow chart of a method for constructing a pathophysiological isomerization graph provided by an embodiment of the present application. Figure 1 As shown, the method includes steps S310 to S330.

[0123] In one feasible embodiment, step S140: fusing multi-dimensional pathological factor information with physiological spatiotemporal information to construct a pathophysiological heterogeneous graph includes:

[0124] S310: Based on the drug dosage and frequency data in the drug effect intensity information and combined with the time window length in the treatment stage division information, the drug dosage data is divided into discrete drug effect intensity quantization values ​​according to the time window length, and each time window corresponds to a drug effect intensity quantization value.

[0125] The quantified value of drug effect intensity refers to a standardized value generated by dividing the drug dosage by time window, reflecting the therapeutic efficacy of the drug within a specific time period. Based on the time window length in the treatment stage division information, such as the hour-level window in the acute phase and the day-level window in the recovery phase, the drug dosage data is divided into discrete quantitative values ​​by time window. For example, in the acute phase of treatment, if the time window is 5 hours, the drug dosage within each 5-hour period is accumulated and multiplied by the baseline intensity coefficient corresponding to the drug type to generate the quantified value of the drug effect intensity for that window.

[0126] S320: Extract infection type data from the infection spread risk information, and map the infection type to infection spread risk level values ​​of different levels based on the change trend of the microcirculation blood flow value within the corresponding time window. The infection spread risk level value increases as the decrease in the microcirculation blood flow value increases.

[0127] The infection spread risk level is dynamically adjusted based on the infection type and microcirculatory blood flow trends, representing the potential threat of infection spread. The infection spread risk level is dynamically adjusted based on infection type data and microcirculatory blood flow trends within the corresponding time window, such as a sustained decrease or a fluctuating increase. If blood flow decreases significantly within the window, the risk level is increased according to a pre-set infection type-risk level mapping table; if blood flow is stable or increasing, the risk level is decreased.

[0128] S330: Arrange the microcirculation blood flow values ​​in the physiological spatiotemporal information in chronological order, and calculate the fluctuation differences between adjacent time points to obtain a microcirculation blood flow fluctuation value sequence.

[0129] Microcirculatory blood flow fluctuation value sequences are time-series data generated by calculating the difference between blood flow values ​​at adjacent time points. These sequences are used to describe the fluctuation characteristics of blood flow dynamics. Microcirculatory blood flow values ​​in physiological spatiotemporal information are arranged chronologically and the differences between adjacent time points are calculated, such as the value at time t+1 minus the value at time t, to generate a fluctuation value sequence. This sequence reflects the magnitude and direction of instantaneous changes in blood flow.

[0130] S340: Establish an association relationship between the quantitative value of the drug effect intensity and the corresponding time window in the microcirculatory blood flow fluctuation value sequence. Each association relationship constitutes an edge of the pathophysiological heterogeneous graph. The weight of the edge is determined by counting the ratio of the direction of drug dose change and the direction of blood flow fluctuation in the same time window. The higher the ratio, the greater the weight of the edge.

[0131] The drug effect intensity quantification values ​​within each time window are time-series aligned with the microcirculatory blood flow fluctuation values. The consistency ratio between the direction of drug dose change and the direction of blood flow fluctuation is calculated. For example, an increase or decrease in drug dose corresponds to an increase or decrease in the direction of blood flow fluctuation. For example, if 80% of the dose increase events within the window are accompanied by an increase in blood flow, an edge with a weight of 0.8 is generated, connecting the drug effect intensity node and the blood flow fluctuation node.

[0132] For example, consider a burn patient. The treatment phase is divided into the acute phase, with a time window of 5 hours. Medication history indicates 2g of intravenous cephalosporin antibiotics every 5 hours. Based on a baseline intensity coefficient of 0.9, the quantified drug effect intensity for each window is 0.9 × 2 = 1.8. The infection type is Staphylococcus aureus, corresponding to an initial risk level of 3. If blood flow decreases by 20% within the window, the risk level increases to 4. Microcirculatory blood flow values ​​are collected hourly, and the differences between adjacent time points are calculated to generate a fluctuation sequence, such as +10 ml / min and -5 ml / min. If the consistency between drug dose increase events and blood flow increase events within the time window is 75%, an edge with a weight of 0.75 is constructed. The resulting pathophysiological heterogeneous graph contains nodes (drug intensity 1.8, risk level 4, fluctuation value +10) and edges (weight 0.75), which are used for subsequent spatiotemporal evolution matrix construction and healing status assessment.

[0133] Based on the same concept, the embodiment of the present application provides a wound healing status assessment system based on artificial intelligence. Figure 4 The artificial intelligence-based wound healing status assessment system provided in the embodiments of the present application is described in detail.

[0134] Figure 4 This is a structural block diagram of an artificial intelligence-based wound healing status assessment system shown in an embodiment of the present application.

[0135] like Figure 4 As shown, the artificial intelligence-based wound healing status assessment system may include:

[0136] Acquisition module 410 is used to acquire the target subject's medical history data and physiological parameter data of the wound area. The medical history data includes treatment time, medication history, and infection type. The physiological parameter data includes hemoglobin oxygenation status values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at multiple time points.

[0137] Extraction module 420 is used to extract multi-dimensional pathological factor information related to wound pathological changes based on the correlation between medication history and infection type in the medical record data. The multi-dimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division;

[0138] A generation module 430 is used to perform cross-time correlation analysis on hemoglobin oxygenation state values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at different time points in the physiological parameter data to generate physiological spatiotemporal information;

[0139] Construction module 440 is used to fuse multi-dimensional pathological factor information with physiological spatiotemporal information to construct a pathophysiological heterogeneous graph. The nodes of the pathophysiological heterogeneous graph include quantified values ​​of drug effect intensity, infection spread risk level values, and microcirculatory blood flow fluctuation values ​​in physiological spatiotemporal information. The edges of the pathophysiological heterogeneous graph represent the correlation between drug effect intensity and microcirculatory blood flow fluctuation values. The weights of the edges are determined based on the temporal matching degree between drug dosage and blood flow changes.

[0140] The generation module 430 is further configured to periodically align the treatment time with the collection time of the physiological parameter data based on the weights of the nodes and edges of the pathophysiological heterogeneous graph to generate a spatiotemporal evolution matrix;

[0141] The generation module 430 is also used to generate wound healing status assessment information based on the correlation between the quantitative value of the drug effect intensity and the microcirculation blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value.

[0142] In one embodiment, the generation module 430 is specifically used to determine the alignment time window length corresponding to the treatment stage based on the treatment stage division information in the multidimensional pathological factor information, wherein the acute treatment stage adopts the first preset time length as the alignment time window length, and the recovery treatment stage adopts the second preset time length as the alignment time window length, and the first preset time length is less than the second preset time length; based on the weight distribution of nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are aligned according to the alignment time window length to generate a spatiotemporal evolution matrix.

[0143] In one embodiment, the generation module 430 is specifically used to divide the treatment time into multiple continuous time segments based on the length of the aligned time window, each time segment contains the quantitative value of the drug effect intensity, the infection spread risk level value and the microcirculation blood flow fluctuation value of the corresponding stage in the treatment stage division information; for the microcirculation blood flow fluctuation value in each time segment, according to the weight distribution of the edges in the pathophysiological heterogeneous graph, the quantitative value of the drug effect intensity connected by the edges with weight values ​​higher than the preset weight threshold is combined with the maximum blood flow fluctuation value in the corresponding time segment to form a time segment dominant parameter group; the quantitative value of the drug effect intensity and the infection spread risk level value in the time segment dominant parameter group are sorted according to the frequency of occurrence within the length of the aligned time window to generate a parameter priority sequence in each time segment; according to the arrangement order of the quantitative value of the drug effect intensity, the infection spread risk level value and the microcirculation blood flow fluctuation value in the parameter priority sequence, the collection time point of the physiological parameter data is mapped to the time point of the treatment time to form a spatiotemporal evolution matrix with time segments as rows and parameter priorities as columns.

[0144] In one embodiment, the generation module 430 is specifically used to divide the wound area into a central area, a marginal area and a surrounding healthy area according to the spatial distribution characteristics of the tissue metabolic heat distribution values ​​in the spatiotemporal evolution matrix; according to the correlation between the quantitative value of the drug action intensity and the microcirculation blood flow fluctuation value in the spatiotemporal evolution matrix, the correlation strength value between the quantitative value of the drug action intensity and the microcirculation blood flow fluctuation value in the central area, the marginal area and the surrounding healthy area is calculated respectively; based on the difference in the correlation strength value in the central area, the marginal area and the surrounding healthy area, combined with the spatial distribution characteristics of the infection spread risk level value, the wound healing status assessment information of the wound area is generated.

[0145] In one embodiment, the generation module 430 is specifically used to superimpose the association strength value of the central area with the data of the corresponding spatial position in the infection spread risk level value to generate a healing activity parameter of the central area, and at the same time, weightedly fuse the association strength value of the edge area with the infection spread risk level value according to the inverse of the time window length in the treatment stage division information to generate a healing inhibition parameter of the edge area; calculate the numerical difference between the healing activity parameter and the healing inhibition parameter within the time window length, and when the growth of the healing activity parameter in the continuous time window is greater than the growth of the healing inhibition parameter, mark the healing status of the central area and the edge area as a positive difference mode; match the positive difference mode with the distribution gradient direction of the infection spread risk level value from the central area to the surrounding healthy area, and determine the wound healing status evaluation information of the wound area according to the gradient direction of the infection spread risk level value and the growth direction of the healing activity parameter in the positive difference mode, and the wound healing status evaluation information includes the tissue regeneration stage and the inflammation retention stage.

[0146] In one embodiment, the extraction module 420 is specifically used to calculate the single medication intensity value based on the drug type and single medication dosage data in the medication history, combined with the preset correspondence between the drug type and the benchmark intensity, to generate drug action intensity information; determine the infection spread risk level value based on the pathogen type corresponding to the infection type, and the preset correspondence between the pathogen type and the infection spread level; determine the treatment stage division information based on the medication frequency in the medication history, and the treatment stage division information includes the recovery treatment stage and the acute treatment stage.

[0147] In one embodiment, the construction module 440 is specifically configured to segment the drug dosage data into discrete drug dosage intensity quantified values ​​according to the time window length based on the drug dosage and frequency data in the drug dosage intensity information and the time window length in the treatment stage division information, with each time window corresponding to a drug dosage intensity quantified value; extract infection type data from the infection spread risk information, and map the infection type to different levels of infection spread risk level values ​​based on the changing trend of the microcirculation blood flow value within the corresponding time window, with the infection spread risk level value increasing as the decrease in the microcirculation blood flow value increases; arrange the microcirculation blood flow values ​​in the physiological spatiotemporal information in chronological order, calculate the fluctuation difference between adjacent time points to obtain a microcirculation blood flow fluctuation value sequence; and establish an association relationship between the drug dosage quantified value and the corresponding time window in the microcirculation blood flow fluctuation value sequence, with each association relationship constituting an edge in the pathophysiological heterogeneous graph. The weight of the edge is determined by statistically analyzing the ratio of the drug dosage change direction to the blood flow fluctuation direction within the same time window. The higher the ratio, the greater the edge weight.

[0148] Figure 4 Each module in the system shown has the function of implementing Figures 1 to 3 The functions of each step in the embodiment can achieve the corresponding technical effects, which will not be described in detail here for the sake of brevity.

[0149] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in one embodiment of the present application is shown.

[0150] The electronic device may include a processor 510 and a memory 520 storing computer program instructions.

[0151] Specifically, the processor 510 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0152] The memory 520 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 520 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 520 may include removable or non-removable (or fixed) media. Where appropriate, the memory 520 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 520 is a non-volatile solid-state memory.

[0153] The memory may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Thus, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to the first aspect of the present disclosure.

[0154] The processor 510 reads and executes computer program instructions stored in the memory 520 to implement any one of the artificial intelligence-based wound healing status assessment methods in the above embodiments.

[0155] In one example, the electronic device may further include a communication interface 530 and a bus 540. Figure 5 As shown, the processor 510 , the memory 520 , and the communication interface 530 are connected via a bus 540 and communicate with each other.

[0156] The communication interface 530 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0157] Bus 540 includes hardware, software or both, and the parts of online data flow metering equipment are coupled to each other. For example, but not limitation, bus can include accelerated graphics port (AGP) or other graphics bus, enhanced industry standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industry standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 540 can include one or more buses. Although the present application embodiment describes and shows specific bus, the application considers any suitable bus or interconnection.

[0158] The electronic device can execute the wound healing status assessment method based on artificial intelligence in the embodiment of the present application, thereby realizing the combination of Figures 1 to 3 Described is an artificial intelligence-based wound healing status assessment method.

[0159] In addition, in conjunction with the artificial intelligence-based wound healing status assessment method in the above-mentioned embodiments, embodiments of the present application may provide a computer-readable storage medium for implementation. The computer-readable storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any of the artificial intelligence-based wound healing status assessment methods in the above-mentioned embodiments is implemented.

[0160] It should be understood that the present application is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and illustrated as examples. However, the method process of the present application is not limited to the specific steps described and illustrated. Those skilled in the art can make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present application.

[0161] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium that can store or transmit information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROMs, flash memories, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0162] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps. In other words, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0163] Aspects of the present application have been described above with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine so that these instructions executed via the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. This processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor or a field programmable logic circuit. It is also understood that each box in the block diagram and / or the flowchart and the combination of the boxes in the block diagram and / or the flowchart can also be implemented by the dedicated hardware that performs the specified function or action, or can be implemented by the combination of dedicated hardware and computer instructions.

[0164] The above description is only a specific embodiment of the present application. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the scope of protection of the present application.

Claims

1. A wound healing status assessment method based on artificial intelligence, characterized in that: include: Obtaining the target subject's medical history data and physiological parameter data of the wound area, wherein the medical history data includes treatment time, medication history, and infection type, and the physiological parameter data includes hemoglobin oxygenation status values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at multiple time points; Extracting multidimensional pathological factor information related to wound pathological changes based on the correlation between medication history and infection type in the medical record data, wherein the multidimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division; Performing cross-time correlation analysis on hemoglobin oxygenation state values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at different time points in the physiological parameter data to generate physiological spatiotemporal information; The multi-dimensional pathological factor information is integrated with the physiological spatiotemporal information to construct a pathophysiological heterogeneous graph, wherein the nodes of the pathophysiological heterogeneous graph include a quantified value of the drug effect intensity, a value of the infection spread risk level, and a microcirculatory blood flow fluctuation value in the physiological spatiotemporal information. The edges of the pathophysiological heterogeneous graph represent the correlation between the drug effect intensity and the microcirculatory blood flow fluctuation value, and the weight of the edge is determined based on the temporal matching degree between the drug dosage and the blood flow change; Based on the weights of the nodes and edges of the pathophysiological heterogeneous graph, the treatment time is periodically aligned with the collection time of the physiological parameter data to generate a spatiotemporal evolution matrix; According to the correlation between the quantitative value of drug action intensity and the microcirculation blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value, wound healing status assessment information is generated.

2. The method according to claim 1, characterized in that The step of periodically aligning the treatment time with the collection time of the physiological parameter data based on the weights of the nodes and edges of the pathophysiological heterogeneous graph to generate a spatiotemporal evolution matrix includes: Determine, based on the treatment stage division information in the multidimensional pathological factor information, the alignment time window length corresponding to the treatment stage, wherein the acute treatment stage uses a first preset time length as the alignment time window length, and the recovery treatment stage uses a second preset time length as the alignment time window length, and the first preset time length is less than the second preset time length; Based on the weight distribution of nodes and edges of the pathophysiological heterogeneous graph, the treatment time and the collection time of the physiological parameter data are aligned according to the length of the alignment time window to generate a spatiotemporal evolution matrix.

3. The method according to claim 2, characterized in that The weight distribution of nodes and edges of the pathophysiological heterogeneous graph is based on aligning the treatment time with the collection time of the physiological parameter data according to the length of the alignment time window to generate a spatiotemporal evolution matrix, including: Dividing the treatment time into a plurality of continuous time segments based on the length of the aligned time window, each time segment comprising a quantitative value of the drug effect intensity, an infection spread risk level value, and the microcirculation blood flow fluctuation value of the corresponding stage in the treatment stage division information; For the microcirculatory blood flow fluctuation value in each time segment, according to the weight distribution of the edges in the pathophysiological heterogeneous graph, the quantified value of the drug effect intensity connected by the edges with weight values ​​higher than a preset weight threshold is combined with the maximum blood flow fluctuation value in the corresponding time segment to form a time segment dominant parameter group; Sort the drug effect intensity quantification values ​​and the infection spread risk level values ​​in the time segment dominant parameter group according to the frequency of occurrence within the alignment time window length to generate a parameter priority sequence within each time segment; According to the arrangement order of the quantitative values ​​of drug effect intensity, infection spread risk level values ​​and microcirculation blood flow fluctuation values ​​in the parameter priority sequence, the collection time points of the physiological parameter data are mapped to the time points of the treatment time to form a spatiotemporal evolution matrix with time segments as rows and the parameter priorities as columns.

4. The method according to claim 1, wherein Based on the correlation between the quantitative value of the drug effect intensity and the microcirculatory blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value, wound healing status assessment information is generated, including: Dividing the wound surface area into a central area, a marginal area, and a surrounding healthy area according to the spatial distribution characteristics of the tissue metabolic heat distribution values ​​in the spatiotemporal evolution matrix; Calculating the correlation strength values ​​between the quantified values ​​of drug action intensity and the microcirculation blood flow fluctuation values ​​in the central area, the edge area, and the surrounding healthy area, respectively, based on the correlation relationship between the quantified values ​​of drug action intensity and the microcirculation blood flow fluctuation values ​​in the spatiotemporal evolution matrix; Based on the differences in the association strength values ​​among the central area, the edge area, and the surrounding healthy area, and combined with the spatial distribution characteristics of the infection spread risk level values, wound healing status assessment information of the wound area is generated.

5. The method according to claim 4, characterized in that Based on the difference in the association strength values ​​among the central area, the edge area, and the surrounding healthy area, combined with the spatial distribution characteristics of the infection spread risk level values, wound healing status assessment information of the wound area is generated, including: The association strength value of the central area is superimposed with the data of the corresponding spatial position in the infection spread risk level value to generate the healing activity parameter of the central area. At the same time, the association strength value of the edge area and the infection spread risk level value are weightedly fused according to the inverse of the time window length in the treatment stage division information to generate the healing inhibition parameter of the edge area. Calculating the numerical difference between the healing activity parameter and the healing inhibition parameter within the time window length, and marking the healing states of the central area and the edge area as a positive difference mode when the increase in the healing activity parameter within the continuous time window is greater than the increase in the healing inhibition parameter; The positive difference pattern is matched with the distribution gradient direction of the infection spread risk level value from the central area to the surrounding healthy area. According to the gradient direction of the infection spread risk level value and the growth direction of the healing activity parameter in the positive difference pattern, the wound healing status assessment information of the wound area is determined. The wound healing status assessment information includes the tissue regeneration stage and the inflammation retention stage.

6. The method according to claim 1, characterized in that The multi-dimensional pathological factor information related to the pathological changes of the wound surface is extracted based on the correlation between the medication history and the infection type in the medical record data. The multi-dimensional pathological factor information includes drug effect intensity information, infection spread risk information and treatment stage division information, including: Calculating a single medication intensity value based on the medication type and single medication dosage data in the medication history, combined with a preset correspondence between the medication type and the reference intensity, to generate the medication intensity information; Determining the infection spread risk level value based on the pathogen type corresponding to the infection type and a preset correspondence between the pathogen type and the infection spread level; The treatment stage division information is determined according to the medication frequency in the medication history, and the treatment stage division information includes a recovery treatment stage and an acute treatment stage.

7. The method according to claim 1, characterized in that The step of fusing the multi-dimensional pathological factor information with the physiological spatiotemporal information to construct a pathophysiological heterogeneous graph includes: Based on the medication dosage and medication frequency data in the medication effect intensity information, combined with the time window length in the treatment stage division information, the medication dosage data is divided into discrete medication effect intensity quantified values ​​according to the time window length, and each time window corresponds to a medication effect intensity quantified value; Extracting infection type data from the infection spread risk information, and mapping the infection type into infection spread risk level values ​​of different levels based on the change trend of the microcirculation blood flow value within the corresponding time window, wherein the infection spread risk level value increases as the decrease in the microcirculation blood flow value increases; Arranging the microcirculation blood flow values ​​in the physiological spatiotemporal information in chronological order, and calculating the fluctuation differences between adjacent time points to obtain a microcirculation blood flow fluctuation value sequence; An association relationship is established between the quantified value of the drug effect intensity and the corresponding time window in the microcirculation blood flow fluctuation value sequence, each association relationship constitutes an edge of the pathophysiological heterogeneous graph, and the weight of the edge is determined by counting the ratio of the direction of drug dose change and the direction of blood flow fluctuation in the same time window. The higher the ratio, the greater the weight of the edge.

8. An artificial intelligence-based wound healing status assessment system, characterized in that: The system comprises: An acquisition module is used to obtain the target subject's medical history data and physiological parameter data of the wound area. The medical history data includes treatment time, medication history, and infection type. The physiological parameter data includes hemoglobin oxygenation status values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at multiple time points. An extraction module is configured to extract multidimensional pathological factor information related to wound pathological changes based on the correlation between medication history and infection type in the medical record data, wherein the multidimensional pathological factor information includes information on drug effect intensity, infection spread risk, and treatment stage division information; a generation module for performing a cross-time correlation analysis on the hemoglobin oxygenation state values, microcirculatory blood flow values, and tissue metabolic heat distribution values ​​at different time points in the physiological parameter data to generate physiological spatiotemporal information; a construction module for fusing the multidimensional pathological factor information with the physiological spatiotemporal information to construct a pathophysiological heterogeneous graph, wherein the nodes of the pathophysiological heterogeneous graph include quantified values ​​of drug action intensity, infection spread risk level values, and microcirculatory blood flow fluctuation values ​​in the physiological spatiotemporal information; the edges of the pathophysiological heterogeneous graph represent the correlation between drug action intensity and microcirculatory blood flow fluctuation values, and the weights of the edges are determined based on the temporal matching degree between drug dosage and blood flow changes; The generation module is further configured to periodically align the treatment time with the collection time of the physiological parameter data based on the weights of the nodes and edges of the pathophysiological heterogeneous graph to generate a spatiotemporal evolution matrix; The generation module is also used to generate wound healing status assessment information based on the correlation between the quantitative value of the drug effect intensity and the microcirculation blood flow fluctuation value in the spatiotemporal evolution matrix, combined with the spatial distribution characteristics of the infection spread risk level value and the tissue metabolic heat distribution value.

9. An electronic device, characterized in that: The device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the artificial intelligence-based wound healing status assessment method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by the processor, the artificial intelligence-based wound healing status assessment method according to any one of claims 1 to 7 is implemented.

Images