Method and system for early prediction of toxic epidermal necrolysis

Abstract

The invention provides a toxic epidermal necrolysis early prediction method and system, and the method comprises the steps: firstly constructing the baseline data information of a patient before the tumor treatment of the patient; and acquiring medical history information and a tumor treatment scheme of the patient. And collecting a skin image of the specific area according to a preset period during follow-up visit. And analyzing the preprocessed skin image of the specific area to obtain pigmentation information and epidermal layer shedding information. And generating skin color change information according to the pigmentation information and the epidermal layer shedding information. And comparing the skin color change information with the baseline data to obtain an abnormal change trend. And inputting the medical history information, the tumor treatment scheme and the abnormal change trend into a preset prediction model, and outputting a toxic epidermal necrosis debonding disease early prediction result of the patient by the prediction model so as to complete the toxic epidermal necrosis debonding disease early prediction of the patient. The method improves the early prediction capability of diseases, and has important significance on prevention and management of toxic epidermal necrosis debonding.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an early prediction method and system for toxic epidermal necrolysis. Background Art

[0002] Toxic epidermal necrolysis (TEN) is a delayed hypersensitivity reaction characterized by mucosal and skin damage. Immunotherapy for malignant tumors (especially immune checkpoint inhibitors) is one of the high-risk factors for TEN. If TEN is not effectively intervened in the early stage, the mortality rate can reach 50%. The main cause of death is systemic epidermal sepsis.

[0003] Due to the late onset of TEN, if injury scoring and subsequent pathogen and drug resistance testing are performed after the patient develops skin lesions, there is a great risk that TEN may rapidly develop sepsis, thrombocytopenia, increased creatinine, or gradually worsen skin lesion symptoms, the area of ​​skin peeling gradually increases, and is accompanied by gastrointestinal mucosal reactions, etc. during the testing process.

[0004] Therefore, there is an urgent need for an early prediction method for toxic epidermal necrolysis that can provide early warning to doctors for patients who already have signs of toxic epidermal necrolysis during the follow-up week.

[0005] Therefore, it is urgent to design a technical solution to solve at least one of the above technical problems. Summary of the invention

[0006] The embodiments of the present application provide a method and system for early prediction of toxic epidermal necrolysis, which can solve the problem that patients who already have signs of toxic epidermal necrolysis during the follow-up week cannot make early predictions of toxic epidermal necrolysis and provide early warnings to doctors.

[0007] In a first aspect, the present application provides an early prediction method for toxic epidermal necrolysis, characterized in that it includes:

[0008] Before the patient undergoes tumor treatment, constructing baseline data information of the patient; the baseline data information includes skin pigment distribution and skin feature information;

[0009] Obtaining the patient's medical history information and tumor treatment plan;

[0010] During the follow-up of the tumor treatment, skin images of a specific area of ​​the patient are collected according to a preset period, and the skin images of the specific area are preprocessed; the preprocessing includes at least noise removal and contrast enhancement;

[0011] Analyzing the preprocessed skin image of the specific area to obtain pigmentation information and epidermal shedding information;

[0012] Generate skin color change information according to the pigment deposition information and epidermal layer shedding information;

[0013] Comparing the skin color change information with the baseline data to obtain abnormal change trends;

[0014] The medical history information, tumor treatment plan and abnormal change trend are input into a preset prediction model, and the prediction model outputs an early prediction result of toxic epidermal necrolysis of the patient, thereby completing the early prediction of toxic epidermal necrolysis of the patient.

[0015] In some embodiments, the capturing of skin images of specific areas of the patient according to a preset cycle includes: obtaining a treatment schedule and a metabolic rate of therapeutic drugs corresponding to the tumor treatment regimen; obtaining an image capture cycle according to the treatment schedule, the metabolic rate of therapeutic drugs and the medical history information; obtaining a specific skin area corresponding to the tumor treatment regimen; the specific skin area includes at least any one of a high-risk area prone to lesions and an area with a previous medical history; capturing images of the specific skin area under standardized capture conditions according to the image capture cycle to obtain skin images of the specific area.

[0016] Exemplarily, the image acquisition cycle is obtained according to the treatment schedule, the metabolic rate of the therapeutic drug and the medical history information, including: obtaining the drug half-life and concentration change function according to the metabolic rate of the therapeutic drug; the concentration change function is used to describe the changes of the drug in the body; obtaining a preset basic acquisition frequency; obtaining the patient's corresponding medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient of a specific high-risk area; the medical history reaction coefficient is used to indicate the patient's sensitivity to the drug, the historical reaction adjustment factor is adjusted according to the patient's previous skin reaction history, and the multi-drug treatment coefficient is used to enhance the interaction of different drugs when the patient receives multiple drug treatments; the image acquisition cycle is calculated according to the drug half-life, concentration change function, basic acquisition frequency, medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient, and the calculation formula of the image acquisition cycle includes:

[0017]

[0018] Wherein, C is the image acquisition cycle, F is the basic acquisition frequency, T d is the half-life of the drug, Cd(t) is the concentration change function, R is the medical history reaction coefficient, H is the historical reaction adjustment factor, E is the environmental impact factor, M is the multi-drug treatment coefficient, A is the age adjustment factor, B is the basic health status coefficient and P is the priority coefficient.

[0019] In some embodiments, the analysis of the preprocessed skin image of the specific area to obtain pigmentation information and epidermal shedding information includes: performing image segmentation on the skin image to obtain multiple skin areas; obtaining relevant areas corresponding to pigmentation and epidermal shedding in the multiple skin areas; obtaining the pigment concentration corresponding to the relevant area, and obtaining the pigmentation information based on the pigment concentration; using a texture analysis tool to detect the epidermal integrity index corresponding to the relevant area; the epidermal integrity index includes any one of texture roughness, texture smoothness, and contrast; obtaining the transparency and reflectivity of the relevant area; and obtaining the epidermal shedding information based on the epidermal integrity index, transparency, and reflectivity.

[0020] Exemplarily, the obtaining of the pigment deposition information according to the pigment concentration includes: performing color space conversion on the relevant area according to the pigment concentration; the color space conversion includes any one of grayscale conversion and Lab space conversion; performing histogram analysis on the relevant area after the color space conversion; performing color histogram equalization on the relevant area after the histogram analysis to obtain the pigment deposition information.

[0021] Exemplarily, the obtaining of the epidermal shedding information according to the epidermal integrity index, transparency and reflectivity includes: respectively obtaining the normal range and abnormal range corresponding to the texture roughness, smoothness, contrast, transparency and reflectivity; respectively generating an epidermal comprehensive evaluation factor according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range; inputting the texture roughness, smoothness, contrast, transparency and reflectivity into a preset multi-factor evaluation model, and the multi-factor evaluation model outputs the multi-factor evaluation result of the patient; and generating the epidermal shedding information according to the epidermal comprehensive evaluation factor and the multi-factor evaluation result.

[0022] It should be noted that, in some embodiments, the generating of the epidermis comprehensive evaluation factor according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively includes: obtaining a first weight corresponding to the texture roughness, a second weight corresponding to the smoothness, a third weight corresponding to the contrast, a fourth weight corresponding to the transparency and a fifth weight corresponding to the reflectivity; wherein the sum of the first weight, the second weight, the third weight, the fourth weight and the fifth weight is 1; standardizing the texture roughness, smoothness, contrast, transparency and reflectivity according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively; calculating the epidermis comprehensive evaluation factor according to the standardized texture roughness, smoothness, contrast, transparency and reflectivity and the first weight, the second weight, the third weight, the fourth weight and the fifth weight; the calculation formula of the epidermis comprehensive evaluation factor includes:

[0023] S=α1*T'1+α2*T'2+α3*T'3+α4*T'4+α5*T'5;

[0024] Among them, S is the comprehensive evaluation factor of the epidermis, α1 is the first weight, α2 is the second weight, α3 is the third weight, α4 is the fourth weight, α5 is the fifth weight, T'1 is the texture roughness, T'2 is the smoothness, T'3 is the contrast, T'4 is the transparency, and T'5 is the reflectivity.

[0025] It should be noted that, in some embodiments, the texture roughness, smoothness, contrast, transparency and reflectivity are standardized according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively, including: determining the maximum and minimum values ​​of the texture roughness, smoothness, contrast, transparency and reflectivity according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively; completing the standardization according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding maximum and minimum values ​​respectively.

[0026] In some embodiments, after completing the early prediction of toxic epidermal necrolysis for the patient, it also includes: obtaining a prediction coefficient corresponding to the early prediction result of toxic epidermal necrolysis; updating the tumor treatment plan according to the prediction coefficient, baseline data information, medical history information and tumor treatment plan; the tumor treatment plan includes a medication plan and a nursing plan.

[0027] In a second aspect, the present application provides an early prediction system for toxic epidermal necrolysis, comprising:

[0028] An information acquisition unit, used to construct baseline data information of the patient before the patient undergoes tumor treatment; the baseline data information includes skin pigment distribution and skin feature information;

[0029] A plan acquisition unit, used to acquire the patient's medical history information and tumor treatment plan;

[0030] An image acquisition unit, used to collect skin images of a specific area of ​​the patient according to a preset period during the follow-up of the tumor treatment, and pre-process the skin images of the specific area; the pre-processing at least includes noise removal and contrast enhancement;

[0031] An image analysis unit, used for analyzing the preprocessed skin image of the specific area to obtain pigmentation information and epidermal shedding information;

[0032] a change generating unit, configured to generate skin color change information according to the pigment deposition information and the epidermal layer shedding information;

[0033] A trend acquisition unit, used to compare the skin color change information with the baseline data to acquire an abnormal change trend;

[0034] The result output unit is used to input the medical history information, tumor treatment plan and abnormal change trend into a preset prediction model, and the prediction model outputs the early prediction result of toxic epidermal necrolysis of the patient to complete the early prediction of toxic epidermal necrolysis of the patient.

[0035] In a third aspect, the present application further provides a computer device comprising a processor and a memory, wherein the memory is used to store a computer program, and when the computer program is executed by the processor, the early prediction method for toxic epidermal necrolysis as described in the first aspect is implemented.

[0036] In a fourth aspect, the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the early prediction method for toxic epidermal necrolysis as described in the first aspect.

[0037] The embodiment of the present application provides a method and system for early prediction of toxic epidermal necrolysis. The method first constructs the baseline data information of the patient before the patient undergoes tumor treatment. The baseline data information includes skin pigment distribution and skin feature information. Then the patient's medical history information and tumor treatment plan are obtained. Then, during the follow-up of tumor treatment, skin images of specific areas of the patient are collected according to a preset period, and the skin images of specific areas are preprocessed; the preprocessing includes at least noise removal and contrast enhancement. Further, the preprocessed skin images of specific areas are parsed to obtain pigment deposition information and epidermal shedding information. Further, skin color change information is generated based on the pigment deposition information and epidermal shedding information. Thus, the skin color change information is compared with the baseline data to obtain the abnormal change trend. Finally, the medical history information, tumor treatment plan and abnormal change trend are input into a preset prediction model, and the prediction model outputs the early prediction results of the patient's toxic epidermal necrolysis, completing the early prediction of the patient's toxic epidermal necrolysis.

[0038] The provided method constructs baseline data information by collecting the patient's skin pigmentation distribution and skin characteristics before the patient undergoes tumor treatment. Obtain the patient's medical history information and specific tumor treatment plan. During the follow-up period, skin images of specific areas are collected according to preset cycles. The images are preprocessed, including noise removal and contrast enhancement, to improve image quality. The preprocessed skin images are parsed to extract pigment deposition information and epidermal shedding information. The skin color change information is compared with the baseline data to extract abnormal change trends. The medical history information, tumor treatment plan and abnormal change trend are input into the preset prediction model. The model outputs the patient's early risk assessment results for toxic epidermal necrolysis.

[0039] The method provided thereby has at least the following beneficial effects:

[0040] 1. Personalization and early prediction: By comparing dynamic skin changes with baseline data, personalized disease risk assessment is achieved, which helps to intervene in the early stages of the disease.

[0041] 2. Automated processing: Automated image analysis and information extraction improve the accuracy and efficiency of detection and reduce human errors.

[0042] 3. Enhanced decision support: Comprehensively consider the patient's medical history, treatment plan and skin change information to provide more accurate and reliable prediction results to help doctors develop more appropriate treatment and management plans.

[0043] 4. Improve patient safety: Through early prediction and detection, timely measures can be taken to reduce the incidence and severity of toxic epidermal necrolysis and improve patient safety and quality of life.

[0044] The provided method improves the early prediction ability of the disease by comprehensively utilizing multifaceted information and automation technology, which is of great significance for the prevention and management of toxic epidermal necrolysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a schematic flow chart of the steps of the early prediction method of toxic epidermal necrolysis shown in an embodiment of the present application;

[0046] Figure 2 This is a schematic diagram of the structure of an early prediction system for toxic epidermal necrolysis shown in an embodiment of the present application;

[0047] Figure 3 A schematic diagram of the structure of a computer device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0048] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, systems, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0049] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0050] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0051] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0052] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0053] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0054] The technical solution of the embodiment of the present application is introduced below.

[0055] Toxic epidermal necrolysis (TEN) is a delayed hypersensitivity reaction characterized by mucosal and skin damage. Immunotherapy for malignant tumors (especially immune checkpoint inhibitors) is one of the high-risk factors for TEN. If TEN is not effectively intervened in the early stage, the mortality rate can reach 50%. The main cause of death is systemic epidermal sepsis.

[0056] Due to the late onset of TEN, if injury scoring and subsequent pathogen and drug resistance testing are performed after the patient develops skin lesions, there is a great risk that TEN may rapidly develop sepsis, thrombocytopenia, increased creatinine, or gradually worsen skin lesion symptoms, the area of ​​skin peeling gradually increases, and is accompanied by gastrointestinal mucosal reactions, etc. during the testing process.

[0057] Therefore, there is an urgent need for an early prediction method for toxic epidermal necrolysis that can provide early warning to doctors for patients who already have signs of toxic epidermal necrolysis during the follow-up week.

[0058] Please refer to Figure 1 , Figure 1 The following is a schematic flow chart of the steps of the early prediction method of toxic epidermal necrolysis shown in the embodiment of the present application. The early prediction method of toxic epidermal necrolysis in the embodiment of the present application can be applied to the computer device provided in any embodiment of the present application. Figure 1 As shown, the early prediction method of toxic epidermal necrolysis in this embodiment includes steps S101 to S107, which are described in detail as follows:

[0059] Step S101, constructing baseline data information of the patient before the patient undergoes tumor treatment; the baseline data information includes skin pigment distribution and skin feature information.

[0060] Specifically, the provided method collects and analyzes the patient's skin data before the patient undergoes tumor treatment, including skin pigment distribution and other skin feature information. Skin pigment distribution uses professional equipment such as a high-definition skin scanner to capture the distribution of pigments on the skin, and record the density, distribution, and differences in different areas of the pigment. Skin feature information may include information such as skin moisture, elasticity, thickness, and temperature. This information can be obtained through skin testing instruments (such as skin analyzers). This information may be obtained through high-definition skin scanners or professional camera equipment and stored in electronic health records. It is also possible to scan the patient's entire body using a digital skin atlas device, record feature information such as pigment distribution and skin texture, and build a patient's skin feature database. This provides the patient with an accurate and stable skin baseline for subsequent monitoring and comparison, thereby improving the accuracy of diagnosis.

[0061] Step S102, obtaining the patient's medical history information and tumor treatment plan.

[0062] Specifically, by collecting the patient's detailed medical history, including past illnesses, allergic reactions, medication use records, etc., as well as current tumor-related treatment information, such as medication type, dosage, and course of treatment. For example, the patient's past medical history and current treatment plan can be retrieved through the electronic medical record system, and information related to skin status can be recorded for subsequent analysis. By providing comprehensive patient background information, it helps identify factors that may affect skin changes or the risk of toxic epidermal necrolysis.

[0063] Step S103, during the follow-up of tumor treatment, skin images of specific areas of the patient are collected according to a preset cycle, and the skin images of specific areas are preprocessed; the preprocessing at least includes noise removal and contrast enhancement.

[0064] Specifically, during the follow-up of tumor treatment, the provided method collects skin images of specific areas of the patient periodically (such as once every 3 days, once a week, or any other period). After image acquisition, preprocessing is performed, including noise removal and contrast enhancement, to ensure image quality. For example, a specific skin area of ​​the patient is regularly photographed using a high-resolution camera every week, and image processing software (such as Photoshop or automated processing tools) is used to remove noise and enhance image contrast. Data consistency is improved through standardized image processing, making subsequent analysis steps more reliable.

[0065] In some embodiments, skin images of specific areas of a patient are acquired according to a preset cycle, including: acquiring a treatment schedule and a metabolic rate of therapeutic drugs corresponding to a tumor treatment regimen; acquiring an image acquisition cycle according to the treatment schedule, the metabolic rate of therapeutic drugs, and medical history information; acquiring a specific skin area corresponding to the tumor treatment regimen; the specific skin area includes at least any one of a high-risk area prone to lesions and an area with a previous medical history; acquiring images of a specific skin area under standardized acquisition conditions according to the image acquisition cycle to acquire skin images of the specific area.

[0066] By obtaining the patient's tumor treatment plan, including the timing of treatment and the type and metabolic characteristics of the drug. This helps to understand the time frame of the drug's action in the body. Combined with the drug's metabolic rate and medical history information, determine when to collect skin images. The goal is to collect when the drug has the greatest impact or skin changes are most obvious. Identify high-risk areas for possible lesions (such as areas exposed to the effects of the drug), including skin areas where the patient has a previous medical history. Collect images of selected areas under preset and consistent conditions (such as using the same camera to shoot under the same lighting conditions to ensure consistency and comparability of each image captured. If conditions permit, use a fixture or positioning marker to ensure that the position, angle, and distance of the camera remain unchanged for each shot) to ensure consistency and comparability of the data.

[0067] By precisely selecting the collection time and area, the chance of capturing treatment-induced skin changes is increased. The collection strategy is adjusted according to the patient's specific treatment and medical history background to improve the accuracy and individual adaptability of monitoring. Standardized collection conditions ensure the reliability of data and the accuracy of subsequent analysis.

[0068] Exemplarily, the image acquisition cycle is obtained according to the treatment schedule, the metabolic rate of the therapeutic drug and the medical history information, including: obtaining the drug half-life and concentration change function according to the metabolic rate of the therapeutic drug; the concentration change function is used to describe the changes of the drug in the body; obtaining the preset basic acquisition frequency; obtaining the patient's corresponding medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient of specific high-risk areas; the medical history reaction coefficient is used to indicate the patient's sensitivity to the drug, the historical reaction adjustment factor is adjusted according to the patient's previous skin reaction history, and the multi-drug treatment coefficient is used to enhance the interaction of different drugs when the patient receives multiple drug treatments; the image acquisition cycle is calculated according to the drug half-life, concentration change function, basic acquisition frequency, medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient, and the calculation formula of the image acquisition cycle includes:

[0069]

[0070] Among them, C is the image acquisition cycle, F is the basic acquisition frequency, T d is the half-life of the drug, C d(t) is the concentration change function, R is the medical history response coefficient, H is the historical response adjustment factor, E is the environmental impact factor, M is the multi-drug treatment coefficient, A is the age adjustment factor, B is the basic health status coefficient and P is the priority coefficient.

[0071] Dynamically adjust the image acquisition frequency according to multiple influencing factors to accurately capture skin changes. Personalized monitoring can be achieved by considering individual differences (such as medical history, age, etc.). Accurately set the acquisition time point to avoid unnecessary frequent acquisition and optimize the use of resources. Through these steps, the efficiency and effectiveness of skin monitoring can be effectively improved, providing solid data support for subsequent pathological analysis and intervention.

[0072] It should be noted that, in some embodiments, if the parameters are:

[0073] F = 3 days; T d = 5 days; C d(t) =0.8 (high drug concentration); R = 0.9 (moderately sensitive reaction); H = 1 (default historical adjustment); E = 1.1; M = 1.2; A = 1; B = 1; P = 1.2. Substituting into the above formula, we get C ≈ 23.7, and the image acquisition cycle is 24 days. This can better adapt to the individual differences of patients, so as to determine a more accurate monitoring cycle and improve the sensitivity and specificity of early detection of toxic epidermal necrolysis.

[0074] Step S104, analyzing the pre-processed skin image of the specific area to obtain pigmentation information and epidermal shedding information.

[0075] Specifically, by analyzing the pre-processed skin images, information about pigmentation and signs of epidermal shedding can be extracted. This step may be accomplished using image recognition algorithms and artificial intelligence technologies. For example, machine learning algorithms (such as convolutional neural networks (CNN)) can be used to analyze processed images to automatically identify and measure skin pigment changes and shedding areas. By monitoring subtle changes in the skin, especially pigmentation and shedding, in real time, it helps to identify abnormal lesions at an early stage.

[0076] Among them, pigmentation information refers to the use of computer vision technology to identify the color information of pixels in the image, identify and quantify pigment changes. Epidermal shedding information refers to analyzing texture changes in the image and capturing the subtle peeling phenomenon of the skin surface, usually relying on texture analysis algorithms.

[0077] In some embodiments, the preprocessed skin image of a specific area is parsed to obtain pigmentation information and epidermal shedding information, including: performing image segmentation on the skin image to obtain multiple skin areas; obtaining relevant areas corresponding to pigmentation and epidermal shedding in the multiple skin areas; obtaining the pigment concentration corresponding to the relevant area, and obtaining pigmentation information based on the pigment concentration; using a texture analysis tool to detect the epidermal integrity index corresponding to the relevant area; the epidermal integrity index includes any one of texture roughness, texture smoothness, and contrast; obtaining the transparency and reflectivity of the relevant area; and obtaining epidermal shedding information based on the epidermal integrity index, transparency, and reflectivity.

[0078] By using image segmentation algorithms (such as K-means clustering, watershed algorithm or convolutional neural network, etc.), the entire skin image is segmented into multiple independent regions so that the characteristics of each region can be finely analyzed. From the multiple segmented regions, those regions related to pigmentation and epidermal shedding are determined. For example, areas with dense pigment or abnormal texture are identified by morphological features or color features. The pigment concentration in the identified area is measured, and pigmentation information is obtained by analyzing the pigment concentration. Texture analysis tools are used to measure the texture roughness, smoothness and contrast of the relevant areas. These indicators can help determine the health of the epidermis. The transparency and reflectivity of the skin are measured by professional image analysis tools, which are also crucial in the assessment of epidermal health. This improves the ability to accurately analyze the skin state and specifically locate possible lesion areas. By comprehensively analyzing multiple features, more accurate information on pigmentation and epidermal shedding can be obtained, providing strong data support for risk assessment. At the same time, by combining advanced image analysis technology, the pigment changes and epidermal state of the skin are deeply analyzed from multiple angles, providing a comprehensive and effective means for the monitoring and early diagnosis of skin lesions.

[0079] Exemplarily, obtaining pigment deposition information according to pigment concentration includes: performing color space conversion on relevant areas according to pigment concentration; color space conversion includes any one of grayscale conversion and Lab space conversion; performing histogram analysis on relevant areas after color space conversion; performing color histogram equalization on relevant areas after histogram analysis to obtain pigment deposition information.

[0080] The sensitivity of the analysis is enhanced by converting the original image to a different color space (such as grayscale space or Lab space). Lab space is particularly helpful for the analysis of color differences. Histogram analysis is performed on the image after color space conversion to identify and quantify the pigment distribution characteristics. The converted histogram is equalized to correct the distribution of pigment concentration and enhance the pigment characteristics to better identify abnormal pigment deposition. In turn, color space conversion and histogram analysis can more clearly distinguish between normal and abnormal pigment distribution. The accuracy of pigment deposition information is improved, providing a more optimized means of skin pigment assessment.

[0081] Exemplarily, epidermal shedding information is obtained based on epidermal integrity indicators, transparency and reflectivity, including: respectively obtaining normal ranges and abnormal ranges corresponding to texture roughness, smoothness, contrast, transparency and reflectivity; generating comprehensive evaluation factors for the epidermis based on texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal ranges and abnormal ranges; inputting texture roughness, smoothness, contrast, transparency and reflectivity into a preset multifactor evaluation model, and the multifactor evaluation model outputs a multifactor evaluation result of the patient; generating epidermal shedding information based on the comprehensive evaluation factors of the epidermis and the multifactor evaluation results.

[0082] Through clinical data or literature, obtain the normal and abnormal ranges of texture roughness, smoothness, contrast, transparency and reflectivity. Evaluate the texture and light reflection characteristics to generate comprehensive evaluation factors of the epidermis. These factors help assess the health of the epidermis. Through the multi-factor evaluation model, various measurement data inputs are comprehensively calculated to output a multi-factor evaluation result reflecting skin health. Comprehensively consider the comprehensive evaluation factors of the epidermis and the multi-factor evaluation results, and finally generate epidermis shedding information. Then, comprehensively consider multiple physical and optical properties to improve the sensitivity and reliability of epidermis shedding detection. The multi-factor evaluation model can more comprehensively reflect complex epidermal changes and provide key information for early intervention.

[0083] It should be noted that, in some embodiments, the epidermis comprehensive evaluation factor is generated according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range, including: obtaining a first weight corresponding to the texture roughness, a second weight corresponding to the smoothness, a third weight corresponding to the contrast, a fourth weight corresponding to the transparency and a fifth weight corresponding to the reflectivity; wherein the sum of the first weight, the second weight, the third weight, the fourth weight and the fifth weight is 1; the texture roughness, smoothness, contrast, transparency and reflectivity are standardized according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range; the epidermis comprehensive evaluation factor is calculated according to the standardized texture roughness, smoothness, contrast, transparency and reflectivity and the first weight, the second weight, the third weight, the fourth weight and the fifth weight; the calculation formula of the epidermis comprehensive evaluation factor includes:

[0084] S=α1*T'1+α2*T'2+α3*T'3+α4*T'4+α5*T'5;

[0085] Among them, S is the comprehensive evaluation factor of the epidermis, α1 is the first weight, α2 is the second weight, α3 is the third weight, α4 is the fourth weight, α5 is the fifth weight, T'1 is texture roughness, T'2 is smoothness, T'3 is contrast, T'4 is transparency, and T'5 is reflectivity.

[0086] By assigning weights to each indicator (texture roughness, smoothness, contrast, transparency, and reflectivity), the sum of these weights is 1. These weights are used to represent the relative importance of each indicator in the assessment of epidermal health. At the same time, by standardizing the measured values ​​of each indicator (texture roughness, smoothness, contrast, transparency, and reflectivity) to a uniform range, the effects of different units or scales can be eliminated. Then, by combining the weights of multiple important indicators, an objective scale of overall health is provided through comprehensive evaluation factors. Converting complex multi-dimensional data into simple and comparable quantitative values ​​helps to quickly identify potential skin problems. The weights can be adjusted according to specific cases to better reflect the differences in epidermal characteristics of different patients.

[0087] It should be noted that, in some embodiments, the texture roughness, smoothness, contrast, transparency and reflectivity are standardized according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal ranges and abnormal ranges, including: determining the maximum and minimum values ​​of the texture roughness, smoothness, contrast, transparency and reflectivity according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal ranges and abnormal ranges; completing the standardization according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding maximum and minimum values.

[0088] For each indicator (texture roughness, smoothness, contrast, transparency and reflectivity), the maximum and minimum values ​​are obtained based on the data of normal and abnormal ranges. Through standardization, indicators of different units or ranges are unified to the same scale to facilitate comprehensive comparison and analysis. At the same time, the deviation of each indicator caused by different measurement scales is removed to improve the accuracy of the calculation of comprehensive evaluation factors. For example, the range of the standardized data is 0 to 1, which more intuitively represents the position of each indicator between normal and abnormal.

[0089] At the same time, the corresponding formulas for standardization include:

[0090]

[0091] Where T std is the standardized value corresponding to each indicator, T max and T min are the maximum and minimum values ​​corresponding to each indicator respectively, and T is the value of each indicator.

[0092] Step S105, generating skin color change information according to the pigmentation information and the epidermal layer shedding information.

[0093] Specifically, by generating skin color change information based on the acquired information on pigmentation and epidermal shedding, this information may be converted into quantifiable indicators, such as skin color change curves or indexes. For example, data analysis tools are used to convert pigmentation and shedding information into change charts, such as using Excel or dedicated analysis software to generate change trend charts. This facilitates quick evaluation and decision-making by providing a visual and quantified overview of skin change trends.

[0094] In some embodiments, pigmentation information is extracted from processed and parsed skin images. This typically involves measuring pigment density and distribution differences on the skin. The pigment data is standardized based on baseline information to eliminate bias between individuals or device acquisition. Standardization methods generally include mean-variance standardization, normalization, etc. Calculate changes in pigment indicators, such as percentage changes compared to baseline pigment or use area-weighted change formulas. Indicators such as pigmentation change rates can be generated. At the same time, by identifying and quantifying the epidermal shedding area in the image, it usually involves in-depth analysis of skin texture or surface features. The percentage change in the shedding area in the image is used to quantify the severity of shedding. This information can help assess the health and recovery of the skin. A well-defined indicator (such as the shedding area change rate) is formed for comparison with further diagnostic criteria. For generating skin color change information based on pigmentation information and epidermal shedding information, the respective indices of pigmentation change and epidermal shedding are combined into a comprehensive skin color change index by applying a weighted formula. Weights may be set based on clinical importance or other statistical analyses, such as the proportion of pigmentation changes to the overall impact of color. At the same time, the trend of skin color change is analyzed by time series methods to identify the speed and direction of change. This usually involves statistical comparison and trend analysis methods. A change curve over time is generated to intuitively show the temporal dynamics of skin pigment and epidermal changes. A heat map or color map is also generated to show the regional location and intensity of the change, intuitively showing which area has the most significant change.

[0095] The provided method then compares the generated comprehensive change information with known clinical benchmarks to determine the normality or abnormality of the skin condition. The generated skin color change information can be provided to clinicians for risk assessment and diagnostic recommendations, such as whether the treatment plan needs to be adjusted.

[0096] Through these steps, complex skin physiological changes can be converted into information that can be directly used for clinical decision-making, improving the understanding and management of the dynamic changes in the patient's skin condition. This data-driven analysis process not only provides quantitative support for skin health monitoring, but also helps prevent the occurrence of serious skin complications.

[0097] Step S106, comparing the skin color change information with the baseline data to obtain abnormal change trends.

[0098] Specifically, by comparing the generated skin color change information with the baseline data, possible abnormal change trends, as well as their severity and development speed can be identified. For example, through abnormal trend analysis, potential signs of skin lesions can be discovered early, so that measures can be taken quickly to prevent the disease from worsening. By statistically comparing the current data with the baseline data, using statistical software such as SPSS or R to analyze abnormal changes, a trend report can be generated.

[0099] In some embodiments, the baseline data is usually skin color information collected in the initial state or healthy state of the patient. The data may include: image data (such as skin photos taken by a high-resolution camera); color space values ​​(such as RGB, HSV or CIE Lab values); normal color difference range and texture information of the skin area. Ensuring that the baseline data is accurate and reliable is the key to subsequent comparative analysis. Real-time skin color data is obtained through regular or continuous monitoring devices (such as handheld devices, mobile phone cameras or dedicated skin scanners, etc.). Preliminary processing is required to eliminate noise and interference (such as changes in lighting). The processing steps may include: Standardization: converting skin color data to the same color space and parameter range. Background removal: focusing on the target skin area to reduce the influence of external factors. Color space conversion and comparative analysis include color space conversion: converting baseline data and real-time data (such as converting to Lab space) to detect color changes more sensitively. Comparative analysis: calculating the difference between two sets of data, usually involving statistical analysis and model algorithms. Specific methods may include: Pixel-level comparison: directly comparing the color differences of corresponding pixels. Regional-level analysis: aggregating the color data of a certain area and comparing it with the baseline. Statistical analysis: Use relevant metrics (e.g., mean, standard deviation) to identify significant changes.

[0100] Abnormal change trend detection includes setting thresholds: setting the threshold between normal and abnormal based on clinical experience or statistical data. Trend analysis: using trend detection algorithms (such as time series analysis, change point detection) to identify gradual or sudden changes in long time periods. Machine learning model: using trained machine learning models to make predictions and identify possible abnormal patterns.

[0101] The method provided can identify abnormal changes in skin color at an early stage and provide early warning for possible pathological changes (such as inflammation, infection, allergic reaction, etc.). It monitors based on patient-specific baseline data and provides customized health management. At the same time, through precise image processing and intelligent analysis, it reduces the interference of environmental factors (light, background) on skin color judgment.

[0102] Step S107, inputting the medical history information, tumor treatment plan and abnormal change trend into a preset prediction model, and the prediction model outputs the early prediction result of the patient's toxic epidermal necrolysis, thereby completing the early prediction of the patient's toxic epidermal necrolysis.

[0103] Specifically, by inputting medical history information, tumor treatment plans, and abnormal change trends into a preset intelligent prediction model, the model combines multiple data and outputs early risk prediction results of toxic epidermal necrolysis through algorithms. For example, machine learning models (such as random forests or deep learning networks) are used to analyze data and output prediction results, and reports are generated based on the prediction results and provided to the medical team for reference. The method uses comprehensive multivariate data for scientific predictions to help doctors make more informed diagnosis and treatment decisions and minimize patient risks.

[0104] Through the combination of the above steps and embodiments, the provided method can effectively improve the accuracy of early prediction of toxic epidermal necrolysis and achieve refined monitoring and management of patient skin health during tumor treatment.

[0105] In some embodiments, after completing the early prediction of toxic epidermal necrolysis for the patient, it also includes: obtaining the prediction coefficient corresponding to the early prediction result of toxic epidermal necrolysis; updating the tumor treatment plan according to the prediction coefficient, baseline data information, medical history information and tumor treatment plan; the tumor treatment plan includes a medication plan and a nursing plan.

[0106] Embodiments extract prediction coefficients from the early prediction results of TEN, which quantify the risk of TEN in patients. It can be obtained based on the output of the machine learning model or the score of the expert system. Collect and integrate the patient's baseline data information (such as age, gender, weight, etc.), medical history information (including previous diseases and treatment history), and the currently used tumor treatment regimen (involving specific drugs and dosages, etc.). According to the risk prediction coefficient, the current medication regimen is adjusted. High risk may mean that some drugs need to be replaced or adjusted to avoid the risk of exacerbating TEN. Pay special attention to drugs that are known to cause TEN, and carefully consider the risks and benefits. Add preventive measures for TEN risks in the care plan to provide personalized care guidance. For example, increase the monitoring frequency for skin problems or adverse reactions. If the patient shows early symptoms, remedial measures can be taken quickly.

[0107] The provided embodiments can better personalize tumor treatment plans and reduce treatment-related risks by comprehensively considering prediction coefficients, patient characteristics and medical history. Timely adjustment of treatment plans can minimize the possibility of TEN and improve the safety of drug treatment. By optimizing medication and nursing plans, the overall efficacy of tumor treatment can be improved to help patients better tolerate treatment. Strengthen the monitoring capabilities for potential serious reactions to ensure that patients have rapid and effective treatment measures when adverse reactions occur. These improvements are not only made with the support of scientific data, but also help improve overall medical quality and patient prognosis by reducing the risk of adverse drug reactions in high-risk patients.

[0108] The method provided in the embodiment of the present application constructs baseline data information by collecting the skin pigment distribution and skin characteristics of the patient before the patient undergoes tumor treatment. Obtain the patient's medical history information and specific tumor treatment plan. During the follow-up period, skin images of specific areas are collected according to a preset period. The image is preprocessed, including noise removal and contrast enhancement, to improve image quality. The preprocessed skin image is parsed to extract pigment deposition information and epidermal shedding information. The skin color change information is compared with the baseline data to extract abnormal change trends. The medical history information, tumor treatment plan and abnormal change trend are input into a preset prediction model. The model outputs the early risk assessment results of toxic epidermal necrolysis in patients.

[0109] The method provided thereby has at least the following beneficial effects:

[0110] 1. Personalization and early prediction: By comparing dynamic skin changes with baseline data, personalized disease risk assessment is achieved, which helps to intervene in the early stages of the disease.

[0111] 2. Automated processing: Automated image analysis and information extraction improve the accuracy and efficiency of detection and reduce human errors.

[0112] 3. Enhanced decision support: Comprehensively consider the patient's medical history, treatment plan and skin change information to provide more accurate and reliable prediction results to help doctors develop more appropriate treatment and management plans.

[0113] 4. Improve patient safety: Through early prediction and detection, timely measures can be taken to reduce the incidence and severity of toxic epidermal necrolysis and improve patient safety and quality of life.

[0114] The provided method improves the early prediction ability of the disease by comprehensively utilizing multifaceted information and automation technology, which is of great significance for the prevention and management of toxic epidermal necrolysis.

[0115] In order to implement the early prediction method of toxic epidermal necrolysis corresponding to the above method embodiment, so as to achieve the corresponding functions and technical effects. Figure 2 , Figure 2 The structural block diagram of a toxic epidermal necrolysis early prediction system 200 provided in an embodiment of the present application is shown. For the sake of convenience, only the parts related to the present embodiment are shown. The toxic epidermal necrolysis early prediction system 200 provided in an embodiment of the present application includes:

[0116] The information acquisition unit 201 is used to construct the baseline data information of the patient before the patient undergoes tumor treatment; the baseline data information includes skin pigment distribution and skin feature information;

[0117] A plan acquisition unit 202 is used to acquire the patient's medical history information and tumor treatment plan;

[0118] An image acquisition unit 203 is used to collect skin images of a specific area of ​​the patient according to a preset period during the follow-up of the tumor treatment, and pre-process the skin images of the specific area; the pre-processing includes at least noise removal and contrast enhancement;

[0119] An image analysis unit 204 is used to analyze the pre-processed skin image of the specific area to obtain pigmentation information and epidermal shedding information;

[0120] A change generating unit 205, configured to generate skin color change information according to the pigment deposition information and the epidermal layer shedding information;

[0121] A trend acquisition unit 206, configured to obtain an abnormal change trend according to the skin color change information and the baseline data;

[0122] The result output unit 207 is used to input the medical history information, tumor treatment plan and abnormal change trend into a preset prediction model, and the prediction model outputs the early prediction result of toxic epidermal necrolysis of the patient to complete the early prediction of toxic epidermal necrolysis of the patient.

[0123] In some embodiments, the image acquisition unit 203 is also used to: obtain the treatment schedule and the metabolic rate of therapeutic drugs corresponding to the tumor treatment plan; obtain the image acquisition cycle according to the treatment schedule, the metabolic rate of therapeutic drugs and the medical history information; obtain the specific skin area corresponding to the tumor treatment plan; the specific skin area includes at least any one of a high-risk area prone to lesions and a previous medical history area; perform image acquisition on the specific skin area under standardized acquisition conditions according to the image acquisition cycle to obtain a skin image of the specific area.

[0124] Exemplarily, the image acquisition unit 203 is also used to: obtain the drug half-life and concentration change function according to the metabolic rate of the therapeutic drug; the concentration change function is used to describe the change of the drug in the body; obtain a preset basic acquisition frequency; obtain the patient's corresponding medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient of a specific high-risk area; the medical history reaction coefficient is used to indicate the patient's sensitivity to the drug, the historical reaction adjustment factor is adjusted according to the patient's previous skin reaction history, and the multi-drug treatment coefficient is used to enhance the interaction of different drugs when the patient receives multiple drug treatments; the image acquisition period is calculated according to the drug half-life, concentration change function, basic acquisition frequency, medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient, and the calculation formula of the image acquisition period includes:

[0125]

[0126] Wherein, C is the image acquisition cycle, F is the basic acquisition frequency, T d is the half-life of the drug, C d(t) is the concentration change function, R is the medical history reaction coefficient, H is the historical reaction adjustment factor, E is the environmental impact factor, M is the multi-drug treatment coefficient, A is the age adjustment factor, B is the basic health status coefficient and P is the priority coefficient.

[0127] In some embodiments, the image parsing unit 204 is further used to: perform image segmentation on the skin image to obtain multiple skin areas; obtain relevant areas corresponding to pigmentation and epidermal shedding in the multiple skin areas; obtain the pigment concentration corresponding to the relevant areas, and obtain the pigmentation information based on the pigment concentration; use a texture analysis tool to detect the epidermal integrity index corresponding to the relevant areas; the epidermal integrity index includes any one of texture roughness, texture smoothness, and contrast; obtain the transparency and reflectivity of the relevant areas; and obtain the epidermal shedding information based on the epidermal integrity index, transparency, and reflectivity.

[0128] Exemplarily, the image analysis unit 204 is also used to: perform color space conversion on the relevant area according to the pigment concentration; the color space conversion includes any one of grayscale conversion and Lab space conversion; perform histogram analysis on the relevant area after color space conversion; perform color histogram equalization on the relevant area after the histogram analysis to obtain the pigment deposition information.

[0129] Exemplarily, the image analysis unit 204 is also used to: respectively obtain the normal range and abnormal range corresponding to the texture roughness, smoothness, contrast, transparency and reflectivity; respectively generate comprehensive evaluation factors of the epidermis according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range; input the texture roughness, smoothness, contrast, transparency and reflectivity into a preset multi-factor evaluation model, and the multi-factor evaluation model outputs the multi-factor evaluation results of the patient; generate the epidermis shedding information according to the comprehensive evaluation factors of the epidermis and the multi-factor evaluation results.

[0130] It should be noted that, in some embodiments, the image analysis unit 204 is also used to: obtain a first weight corresponding to the texture roughness, a second weight corresponding to the smoothness, a third weight corresponding to the contrast, a fourth weight corresponding to the transparency, and a fifth weight corresponding to the reflectivity; wherein the sum of the first weight, the second weight, the third weight, the fourth weight, and the fifth weight is 1; standardize the texture roughness, smoothness, contrast, transparency, and reflectivity according to the texture roughness, smoothness, contrast, transparency, and reflectivity and the corresponding normal range and abnormal range respectively; calculate the epidermis comprehensive evaluation factor according to the standardized texture roughness, smoothness, contrast, transparency, and reflectivity and the first weight, the second weight, the third weight, the fourth weight, and the fifth weight; the calculation formula of the epidermis comprehensive evaluation factor includes:

[0131] S=α1*T'1+α2*T'2+α3*T'3+α4*T'4+α5*T'5;

[0132] Among them, S is the comprehensive evaluation factor of the epidermis, α1 is the first weight, α2 is the second weight, α3 is the third weight, α4 is the fourth weight, α5 is the fifth weight, T'1 is the texture roughness, T'2 is the smoothness, T'3 is the contrast, T'4 is the transparency, and T'5 is the reflectivity.

[0133] It should be noted that, in some embodiments, the image analysis unit 204 is also used to: determine the maximum and minimum values ​​of the texture roughness, smoothness, contrast, transparency and reflectivity according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively; and complete the standardization according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding maximum and minimum values ​​respectively.

[0134] In some embodiments, after completing the early prediction of toxic epidermal necrolysis for the patient, it also includes: a plan generation unit, used to obtain a prediction coefficient corresponding to the early prediction result of toxic epidermal necrolysis; updating the tumor treatment plan according to the prediction coefficient, baseline data information, medical history information and tumor treatment plan; the tumor treatment plan includes a medication plan and a nursing plan.

[0135] The above-mentioned toxic epidermal necrolysis early prediction system 200 can implement the toxic epidermal necrolysis early prediction method of the above-mentioned method embodiment. The options in the above-mentioned method embodiment are also applicable to this embodiment and will not be described in detail here. The rest of the contents of the embodiment of the present application can refer to the contents of the above-mentioned method embodiment, and will not be repeated in this embodiment.

[0136] Figure 3 This is a schematic diagram of the structure of a computer device provided in one embodiment of the present application. Figure 3 As shown, the computer device 3 of this embodiment includes: at least one processor 30 ( Figure 3 Only one is shown in the figure), a memory 31, and a computer program 32 stored in the memory 31 and executable on the at least one processor 30, wherein the processor 30 implements the steps of any of the above method embodiments when executing the computer program 32.

[0137] The computer device 3 may be a computing device such as a smart phone, a tablet computer, a desktop computer, a cloud server, etc. The computer device may include but is not limited to a processor 30 and a memory 31. Those skilled in the art will understand that Figure 3 It is only an example of computer device 3 and does not constitute a limitation on computer device 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components, for example, it may also include input and output devices, network access devices, etc.

[0138] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0139] In some embodiments, the memory 31 may be an internal storage unit of the computer device 3, such as a hard disk or memory of the computer device 3. In other embodiments, the memory 31 may also be an external storage device of the computer device 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 3. Further, the memory 31 may also include both an internal storage unit and an external storage device of the computer device 3. The memory 31 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program. The memory 31 may also be used to temporarily store data that has been output or is to be output.

[0140] In addition, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0141] An embodiment of the present application provides a computer program product. When the computer program product is run on a computer device, the computer device implements the steps in the above-mentioned method embodiments when executing the computer device.

[0142] In several embodiments provided in the present application, it is understood that each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved.

[0143] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0144] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for early prediction of toxic epidermal necrolysis, characterized in that: include: Before the patient undergoes tumor treatment, construct baseline data information of the patient; The baseline data information includes skin pigment distribution and skin feature information; Obtaining the patient's medical history information and tumor treatment plan; During the follow-up of the tumor treatment, skin images of a specific area of ​​the patient are collected according to a preset period, and the skin images of the specific area are preprocessed; the preprocessing includes at least noise removal and contrast enhancement; Analyzing the preprocessed skin image of the specific area to obtain pigmentation information and epidermal shedding information; Generate skin color change information according to the pigment deposition information and epidermal layer shedding information; Comparing the skin color change information with the baseline data to obtain abnormal change trends; The medical history information, tumor treatment plan and abnormal change trend are input into a preset prediction model, and the prediction model outputs an early prediction result of toxic epidermal necrolysis of the patient, thereby completing the early prediction of toxic epidermal necrolysis of the patient.

2. The method according to claim 1, characterized in that The step of acquiring the skin image of a specific area of ​​the patient according to a preset period includes: Obtaining the treatment schedule and therapeutic drug metabolism rate corresponding to the tumor treatment plan; Acquiring an image acquisition cycle according to the treatment schedule, the metabolic rate of the therapeutic drug, and the medical history information; Acquire a specific skin area corresponding to the tumor treatment plan; the specific skin area includes at least one of a high-risk area prone to lesions and an area with a previous medical history; Image acquisition is performed on the specific skin area under standardized acquisition conditions according to the image acquisition cycle to obtain the skin image of the specific area.

3. The method according to claim 2, characterized in that The acquiring of the image acquisition cycle according to the treatment schedule, the metabolic rate of the therapeutic drug and the medical history information comprises: Obtaining a drug half-life and a concentration change function according to the metabolic rate of the therapeutic drug; the concentration change function is used to describe changes in the drug in the body; Get the preset basic acquisition frequency; Obtaining the patient's corresponding medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health condition coefficient and priority coefficient of specific high-risk areas; the medical history reaction coefficient is used to indicate the patient's sensitivity to drugs, the historical reaction adjustment factor is adjusted according to the patient's previous skin reaction history, and the multi-drug treatment coefficient is used to enhance the interactive effects of different drugs when the patient receives multiple drug treatments; The image acquisition cycle is calculated according to the drug half-life, concentration change function, basic acquisition frequency, medical history reaction coefficient, historical reaction adjustment factor, environmental impact factor, multi-drug treatment coefficient, age adjustment factor, basic health status coefficient and priority coefficient. The calculation formula of the image acquisition cycle includes: Wherein, C is the image acquisition cycle, F is the basic acquisition frequency, T d is the half-life of the drug, C d(t) is the concentration change function, R is the medical history reaction coefficient, H is the historical reaction adjustment factor, E is the environmental impact factor, M is the multi-drug treatment coefficient, A is the age adjustment factor, B is the basic health status coefficient and P is the priority coefficient.

4. The method according to claim 1, characterized in that: The analyzing the preprocessed skin image of the specific area to obtain pigmentation information and epidermal shedding information includes: Performing image segmentation on the skin image to obtain multiple skin areas; Acquire relevant areas corresponding to pigmentation and epidermal shedding in the plurality of skin areas; Acquire the pigment concentration corresponding to the relevant area, and acquire the pigment deposition information according to the pigment concentration; Using a texture analysis tool to detect the epidermis integrity index corresponding to the relevant area; the epidermis integrity index includes any one of texture roughness, texture smoothness, and contrast; Obtaining the transparency and reflectivity of the relevant area; The epidermal layer shedding information is obtained according to the epidermal layer integrity index, transparency and reflectivity.

5. The method according to claim 4, characterized in that The step of obtaining the pigment deposition information according to the pigment concentration includes: Performing color space conversion on the relevant area according to the pigment concentration; the color space conversion includes any one of grayscale conversion and Lab space conversion; Performing a histogram analysis on the relevant area after the color space conversion; Color histogram equalization is performed on the relevant area after the histogram analysis to obtain the pigment deposition information.

6. The method according to claim 4, characterized in that The step of obtaining the epidermis shedding information according to the epidermis integrity index, transparency and reflectivity includes: Respectively obtaining normal ranges and abnormal ranges corresponding to the texture roughness, smoothness, contrast, transparency and reflectivity; Generate a comprehensive evaluation factor of the epidermis layer according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively; Inputting the texture roughness, smoothness, contrast, transparency and reflectivity into a preset multi-factor evaluation model, and the multi-factor evaluation model outputs a multi-factor evaluation result of the patient; The epidermal layer shedding information is generated according to the epidermal layer comprehensive evaluation factors and the multi-factor evaluation results.

7. The method according to claim 6, characterized in that The epidermis comprehensive evaluation factors are generated according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range, respectively, including: Obtaining a first weight corresponding to the texture roughness, a second weight corresponding to the smoothness, a third weight corresponding to the contrast, a fourth weight corresponding to the transparency, and a fifth weight corresponding to the reflectivity; wherein the sum of the first weight, the second weight, the third weight, the fourth weight, and the fifth weight is 1; The texture roughness, smoothness, contrast, transparency and reflectivity are standardized according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively: The epidermis comprehensive evaluation factor is calculated according to the standardized texture roughness, smoothness, contrast, transparency and reflectivity and the first weight, the second weight, the third weight, the fourth weight and the fifth weight; the calculation formula of the epidermis comprehensive evaluation factor includes: S=α1*T′1+α2*T′2+α3*T′3+α4*T′4+α5*T′5; Among them, S is the comprehensive evaluation factor of the epidermis, α1 is the first weight, α2 is the second weight, α3 is the third weight, α4 is the fourth weight, α5 is the fifth weight, T′1 is the texture roughness, T′2 is the smoothness, T′3 is the contrast, T′4 is the transparency, and T′5 is the reflectivity.

8. The method according to claim 7, characterized in that The step of standardizing the texture roughness, smoothness, contrast, transparency and reflectivity according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively includes: Determining the maximum and minimum values ​​of the texture roughness, smoothness, contrast, transparency and reflectivity according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding normal range and abnormal range respectively; The normalization is performed according to the texture roughness, smoothness, contrast, transparency and reflectivity and the corresponding maximum and minimum values ​​respectively.

9. The method according to claim 1, characterized in that: After completing the early prediction of toxic epidermal necrolysis for the patient, the method further comprises: Obtaining a prediction coefficient corresponding to the early prediction result of toxic epidermal necrolysis; The tumor treatment plan is updated according to the prediction coefficient, baseline data information, medical history information and tumor treatment plan; the tumor treatment plan includes a medication plan and a nursing plan.

10. An early prediction system for toxic epidermal necrolysis, characterized in that: include: An information acquisition unit, used to construct baseline data information of the patient before the patient undergoes tumor treatment; The baseline data information includes skin pigment distribution and skin feature information; A plan acquisition unit, used to acquire the patient's medical history information and tumor treatment plan; An image acquisition unit, used for collecting skin images of a specific area of ​​the patient according to a preset period during the follow-up of the tumor treatment, and preprocessing the skin images of the specific area; The preprocessing includes at least noise removal and contrast enhancement; An image analysis unit, used for analyzing the preprocessed skin image of the specific area to obtain pigmentation information and epidermal shedding information; a change generating unit, configured to generate skin color change information according to the pigment deposition information and the epidermal layer shedding information; A trend acquisition unit, used to compare the skin color change information with the baseline data to acquire an abnormal change trend; The result output unit is used to input the medical history information, tumor treatment plan and abnormal change trend into a preset prediction model, and the prediction model outputs the early prediction result of toxic epidermal necrolysis of the patient to complete the early prediction of toxic epidermal necrolysis of the patient.

Images