Tumor lysis syndrome evolution risk layering and early warning system

Abstract

The invention relates to the technical field of personal health risk assessment, in particular to a tumor lysis syndrome evolution risk layering and early warning system, which can realize the following steps through mutual cooperation of a plurality of modules: obtaining target metabolic data of different preset metabolic dimensions of a target patient in a current observation time period, a tumor CT image and a peripheral blood smear image are obtained; determining a current cell distribution index according to the cellular morphology condition in the peripheral blood smear image; according to gray level distribution and gradient distribution conditions in the tumor CT image, determining a current structure disintegration possible index; determining a current metabolic diffusion possible index according to the change fluctuation condition of the target metabolic data; and thus, a current risk layered early warning auxiliary value is determined. According to the method, various metabolic data are comprehensively considered, the auxiliary value is quantified in combination with the tumor CT image and the peripheral blood smear image, and the setting reasonability of the auxiliary value is improved, so that the auxiliary effect is improved.

Description

Technical Field

[0001] This invention relates to the field of personal health risk assessment technology, specifically to a tumor lysis syndrome evolution risk stratification and early warning system. Background Technology

[0002] With the development of technology, the application of technologies to assist doctors in risk stratification and early warning is becoming increasingly widespread. For example, it can be used to assist doctors in risk stratification and early warning of tumor lysis syndrome evolution. Currently, the common method used to assist doctors in risk stratification and early warning of tumor lysis syndrome evolution is to use the collected blood uric acid concentration as an auxiliary value for risk stratification and early warning.

[0003] However, when using collected blood uric acid concentrations to assist in risk stratification and early warning of tumor lysis syndrome, the following technical problems often arise: In reality, the risk of tumor lysis syndrome evolution is often influenced by more than just serum uric acid concentration. Therefore, using only the collected serum uric acid concentration to assist in the risk stratification and early warning of tumor lysis syndrome evolution may result in a poorly reasonable final auxiliary value set to help doctors judge the risk of tumor lysis syndrome evolution due to the limited number of factors considered, thus leading to poor auxiliary effect. Summary of the Invention

[0004] To address the technical problem of poor auxiliary effects caused by the unreasonable setting of the final auxiliary value in assisting doctors to judge the risk of tumor lysis syndrome evolution, this invention proposes a tumor lysis syndrome evolution risk stratification and early warning system.

[0005] In a first aspect, the present invention provides a tumor lysis syndrome evolution risk stratification and early warning system, the system comprising: The data acquisition module is used to acquire target metabolic data of the target patient at different preset metabolic dimensions at each acquisition time within the current observation period, and to acquire tumor CT images and peripheral blood smear images of the target patient at the current time, where the current time is the end time of the current observation period. The current cell distribution index determination module is used to determine the current cell distribution index based on the cell morphology in the peripheral blood smear image; The module for determining potential indicators of current structural disintegration is used to determine potential indicators of current structural disintegration based on the grayscale and gradient distribution in tumor CT images. The current metabolic diffusion potential indicator determination module is used to determine the current metabolic diffusion potential indicator based on the changes and fluctuations of target metabolic data of different preset metabolic dimensions at all collection times during the current observation period of the target patient. The current risk stratification early warning auxiliary value determination module is used to determine the current risk stratification early warning auxiliary value based on the current cell distribution index, the current structural disintegration probability index, and the current metabolic diffusion probability index.

[0006] In conjunction with the first aspect above, in one possible implementation, determining the current cell distribution index based on the cell morphology within the peripheral blood smear image includes: The peripheral blood smear image of the target patient at the current time is identified as the target peripheral blood smear image, and cell identification is performed on the target peripheral blood smear image to obtain the target cell region; Based on the actual area of ​​each target cell region, determine the circumference of the regular circle corresponding to each target cell region; The morphological irregularity of each target cell region is determined based on the difference between the regular circular perimeter corresponding to each target cell region and the actual perimeter. The current cell distribution index is determined based on the morphological irregularity of all target cell regions in the target peripheral blood smear image.

[0007] In conjunction with the first aspect above, in one possible implementation, determining the current cell distribution index based on the morphological irregularity corresponding to all target cell regions in the target peripheral blood smear image includes: The target peripheral blood smear image is divided into equal parts to obtain a reference image block; From all target cell regions in the target peripheral blood smear image, select target cell regions whose intersection with the same reference image block is not an empty set, and form the cell region set corresponding to the reference image block; The mean of the morphological irregularity of all target cell regions in the cell region set corresponding to each reference image block is determined as the morphological irregularity representative factor for each reference image block. The current cell distribution index is determined based on the mean of the morphological irregularity corresponding to all target cell regions and the entropy value of the morphological irregularity representative factor corresponding to all reference image patches.

[0008] In conjunction with the first aspect above, in one possible implementation, determining the potential indicators of current structural disintegration based on the grayscale and gradient distribution within the tumor CT image includes: The tumor CT image of the target patient at the current time is identified as the target tumor CT image; The gray value corresponding to each pixel in the target tumor CT image is normalized to obtain the gray normalization factor corresponding to each pixel in the target tumor CT image. The gradient value corresponding to each pixel in the CT image of the target tumor is normalized to obtain the gradient normalization factor corresponding to each pixel in the CT image of the target tumor. Tumor identification is performed on the CT images of the target tumor to obtain the target tumor region; Based on the variance of the gray-level normalization factor corresponding to all pixels within each target tumor region and the mean of the gradient normalization factor corresponding to all pixels on the edge of each target tumor region, determine the possible local disintegration factors for each target tumor region. Based on the potential local disintegration factors corresponding to the target tumor region in the target tumor CT image, determine the potential indicators of current structural disintegration.

[0009] In conjunction with the first aspect above, in one possible implementation, determining the current structural disintegration probability index based on the local disintegration probability factor corresponding to the target tumor region in the target tumor CT image includes: The cumulative value of the local disintegration potential factors corresponding to all target tumor regions in the target tumor CT image is determined as the current structural disintegration potential index.

[0010] In conjunction with the first aspect above, in one possible implementation, determining the current metabolic diffusion potential indicator based on the fluctuations in target metabolic data of different preset metabolic dimensions at all collection times within the current observation period of the target patient includes: Based on the target metabolic data of the target patient at each collection time within the current observation period for different preset metabolic dimensions, determine the metabolic representative factor corresponding to each collection time within the current observation period, wherein the target metabolic data is normalized data. Based on the changes and fluctuations of metabolic representative factors at each collection time and previous collection times within the current observation period, metabolic instability indicators corresponding to each collection time within the current observation period are determined. Based on the metabolic instability indicators corresponding to all collection times within the current observation period, the possible indicators of current metabolic diffusion are determined.

[0011] In conjunction with the first aspect above, in one possible implementation, determining the metabolic representative factor corresponding to each collection time within the current observation period based on the target metabolic data of the target patient at each collection time with different preset metabolic dimensions within the current observation period includes: Any collection time within the current observation period is designated as the marker time, and the mean of all target metabolic data of the target patient in all preset metabolic dimensions at the marker time is designated as the metabolic representative factor corresponding to the marker time.

[0012] In conjunction with the first aspect above, in one possible implementation, determining the metabolic instability index corresponding to each collection moment within the current observation period based on the changes and fluctuations of the metabolic representative factors corresponding to each collection moment and previous collection moments within the current observation period includes: Based on the metabolic representative factors corresponding to all collection times within the current observation period, the instantaneous rate of change and acceleration corresponding to each collection time within the current observation period are determined by the central difference method. Any acquisition time within the current observation period is designated as the marker time; If the instantaneous rate of change corresponding to the marked time is negative, then the target change factor corresponding to the marked time is set to 0; if the instantaneous rate of change corresponding to the marked time is not negative, then the instantaneous rate of change corresponding to the marked time is determined as the target change factor corresponding to the marked time. If the acceleration corresponding to the marked time is negative, then the reference change factor corresponding to the marked time is set to 0; if the acceleration corresponding to the marked time is not negative, then the acceleration corresponding to the marked time is determined as the reference change factor corresponding to the marked time. Based on the entropy values ​​of all representative metabolic factors within the preset window period corresponding to the marked time, as well as the target change factor and reference change factor corresponding to the marked time, the metabolic instability index corresponding to the marked time is determined, wherein the marked time is the end time of the preset window period corresponding to it.

[0013] In conjunction with the first aspect above, in one possible implementation, determining the current metabolic diffusion potential index based on the metabolic instability indexes corresponding to all collection times within the current observation period includes: Divide the current observation time period into equal parts to obtain reference time periods, and obtain a reference time period sequence; The mean value of metabolic instability indicators corresponding to all collection times within each reference period is determined as the representative factor of metabolic instability for each reference period. The difference between the representative factors of metabolic instability corresponding to each adjacent reference time period in the reference time period sequence is determined as the local diffusion factor, thus obtaining the local diffusion factor sequence; Based on the mean of all positive local diffusion factors in the local diffusion factor sequence, the variance of the representative factors of metabolic instability corresponding to all reference time periods, and the mean of metabolic instability indicators corresponding to all collection times within the current observation time period, the possible indicators of current metabolic diffusion are determined.

[0014] In conjunction with the first aspect above, in one possible implementation, determining the current risk stratification early warning auxiliary value based on the current cell distribution index, the current structural disintegration probability index, and the current metabolic diffusion probability index includes: The sum of the current cell distribution index, the current structural disintegration probability index, and the current metabolic diffusion probability index is determined as the current risk stratification early warning auxiliary value.

[0015] Secondly, the present invention provides a method for risk stratification and early warning of tumor lysis syndrome evolution, implemented through a tumor lysis syndrome evolution risk stratification and early warning system, the method comprising: Acquire target metabolic data of the target patient at each collection time within the current observation period for different preset metabolic dimensions, and acquire tumor CT images and peripheral blood smear images of the target patient at the current time, where the current time is the end time of the current observation period; The current cell distribution index is determined based on the cell morphology in the peripheral blood smear image; Based on the grayscale and gradient distribution in tumor CT images, potential indicators of current structural disintegration can be determined. Based on the changes and fluctuations of target metabolic data of different preset metabolic dimensions at all collection times during the current observation period, the possible indicators of current metabolic diffusion are determined. Based on the current cell distribution indicators, current structural disintegration potential indicators, and current metabolic diffusion potential indicators, determine the current risk stratification early warning auxiliary value.

[0016] Thirdly, a server is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, enabling the device to execute the aforementioned method for risk stratification and early warning of tumor lysis syndrome evolution.

[0017] Fourthly, a computer program product is provided, comprising: computer program code, which, when run on a computer, causes the computer to execute the aforementioned method for risk stratification and early warning of tumor lysis syndrome evolution.

[0018] Fifthly, a computer-readable storage medium is provided that stores computer program code, which, when executed on a computer, causes the computer to perform the aforementioned method for risk stratification and early warning of tumor lysis syndrome evolution.

[0019] The present invention has the following beneficial effects: This invention discloses a tumor lysis syndrome evolution risk stratification and early warning system. It comprehensively considers various metabolic data and combines tumor CT images and peripheral blood smear images to quantify auxiliary values. This solves the technical problem of poor auxiliary effect caused by the poor rationality of the final auxiliary value setting in assisting physicians to judge the evolution risk of tumor lysis syndrome, and improves the rationality of auxiliary value setting, thereby improving the auxiliary effect. Specifically, this invention quantifies the current cell distribution index and the current structural disintegration probability index based on cell morphology in peripheral blood smear images and grayscale and gradient distribution in tumor CT images. Furthermore, by analyzing the changes and fluctuations of various metabolic data related to the tumor lysis syndrome evolution risk of the target patient during the current observation period, it quantifies the current metabolic diffusion probability index. This quantifies the current risk stratification and early warning auxiliary value for assisting physicians in judging the evolution risk of tumor lysis syndrome, improving the rationality of auxiliary value setting and thus enhancing the auxiliary effect. Attached Figure Description

[0020] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the structure of a tumor lysis syndrome evolution risk stratification and early warning system according to the present invention; Figure 2 This is a flowchart of a method for risk stratification and early warning of tumor lysis syndrome evolution according to the present invention; Figure 3 This is a schematic diagram of the structure of a computer device according to the present invention. Detailed Implementation

[0022] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the technical solution proposed according to the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] Tumor lysis syndrome (TLS) is a common and potentially life-threatening oncological emergency in medical oncology, hematology, and intensive care units. It typically results from highly effective anticancer therapy or spontaneous tumor remission, leading to the rapid destruction and lysis of a large number of tumor cells. Traditional methods relying on clinical experience or simple scoring often fail to capture individual differences (such as baseline renal function, tumor type, and treatment regimen), resulting in insufficient accuracy in risk stratification. Therefore, a systematic and intelligent approach is often needed for individualized and dynamically updated risk stratification, enabling earlier intervention, minimizing the risk of severe TLS and its complications (especially renal failure), and improving quality of life and prognosis.

[0025] refer to Figure 1 The diagram illustrates a structural schematic of a tumor lysis syndrome evolution risk stratification and early warning system according to the present invention. The tumor lysis syndrome evolution risk stratification and early warning system includes: The data acquisition module 101 is used to acquire target metabolic data of different preset metabolic dimensions of the target patient at each acquisition time within the current observation period, and to acquire tumor CT images and peripheral blood smear images of the target patient at the current time.

[0026] The target patient can be a patient currently suffering from tumor lysis syndrome. The current observation period can characterize the current treatment observation process for tumor lysis syndrome. The start time of the current observation period can be the time when the target patient is diagnosed with tumor lysis syndrome. The end time of the current observation period can be the current time. The duration between adjacent collection times within the current observation period can be preset, and can be set according to the actual scenario, such as 1 day. The preset metabolic dimensions can be preset metabolic dimensions related to the risk of tumor lysis syndrome evolution, and can be set according to the actual scenario. The number of preset metabolic dimensions can be preset, and can be set according to the actual scenario, such as 3. The preset metabolic dimensions can be, but are not limited to: serum uric acid concentration dimension, serum potassium concentration dimension, and serum phosphorus concentration dimension. The target metabolic data can be the normalized value of metabolic data under the preset metabolic dimensions. For example, the metabolic data under the serum uric acid concentration dimension can be serum uric acid concentration. The normalized value of the metabolic data can be calculated using the max-min normalization algorithm or the Z-score normalization formula. For example, the target metabolic data for the serum uric acid concentration dimension can be the normalized value of serum uric acid concentration.

[0027] As an example, this step may include the following steps: The first step is to obtain target metabolic data for the target patient at each collection time within the current observation period, for different preset metabolic dimensions.

[0028] For example, taking blood potassium concentration as an example, at each collection time within the current observation period, the blood potassium concentration of the target patient can be collected by venous blood sampling, and the normalized value of the collected blood potassium concentration is used as the target metabolic data for the blood potassium concentration dimension.

[0029] Alternatively, the method for acquiring metabolic data can also be as follows: the system automatically collects and tracks metabolic data related to the core pathogenesis of tumor lysis syndrome (TLS) through a standardized laboratory information system interface, including blood uric acid concentration, blood potassium concentration, blood phosphorus concentration, etc.

[0030] The second step is to obtain the tumor CT images of the target patient at the current moment.

[0031] For example, at the current moment, a 64-slice spiral CT (Computed Tomography) scanner can be used to scan the primary tumor and metastases of the target patient. The scan range covers the organ containing the tumor and the surrounding 2cm of normal tissue, generating CT tomographic images in DICOM (Digital Imaging and Communications in Medicine) format. These images are then preprocessed and recorded as tumor CT images. The relevant parameters for the 64-slice spiral CT scanner can be set as follows: slice thickness 1mm, reconstruction interval 0.5mm, spatial resolution ≤0.625mm×0.625mm; scan time ≤5 seconds / site. Image preprocessing can include, but is not limited to, image enhancement and image denoising. Image enhancement can be achieved using an adaptive histogram equalization algorithm, primarily used to enhance image contrast. Image denoising can be achieved using Gaussian filtering, primarily used to remove noise interference.

[0032] The third step is to obtain the peripheral blood smear image of the target patient at the current moment.

[0033] For example, at the current moment, a fully automated cell morphology analyzer can be used to scan the peripheral blood smear of the target patient, generating a digital cell morphology image, which is recorded as a peripheral blood smear image. The relevant parameters of the fully automated cell morphology analyzer can be set as follows: optical magnification supports switching between 20× / 40× / 100×; resolution 0.1μm / pixel; cell recognition accuracy ≥98%. Peripheral blood typically refers to circulating blood collected through venous puncture, and usually contains circulating tumor cells, tumor cell lysis fragments, and affected normal blood cells.

[0034] It should be noted that edge enhancement can also be performed using the Sobel operator, which can, to some extent, ensure the accuracy of subsequent cell segmentation.

[0035] The current cell distribution index determination module 102 is used to determine the current cell distribution index based on the cell morphology in the peripheral blood smear image.

[0036] It should be noted that the onset of tumor lysis syndrome often stems from the uncontrolled rupture of tumor cell membranes, which often manifests as irregular morphological features such as cell swelling and chromatin aggregation. When a large number of cells dissolve in a disordered manner, their spatial distribution in blood smears often exhibits a high degree of randomness and disorder.

[0037] As an example, determining the current cell distribution index may include the following steps: The first step is to identify the peripheral blood smear image of the target patient at the current time as the target peripheral blood smear image, and to perform cell identification on the target peripheral blood smear image to obtain the target cell region.

[0038] Among them, the target cell region can characterize the cell.

[0039] For example, cell identification can be performed on target peripheral blood smear images using a U-Net network segmentation model or a threshold segmentation algorithm, and the identified cell regions can be recorded as target cell regions.

[0040] The second step is to determine the circumference of the regular circle corresponding to each target cell region based on the actual area of ​​each target cell region.

[0041] The actual area of ​​the target cell region can be represented by the total number of all pixels in the target cell region.

[0042] For example, the formula for determining the circumference of a regular circle corresponding to a target cell region can be: ; in, The first in the target peripheral blood smear image i The circumference of a regular circle corresponding to each target cell region. i It is the sequence number of the target cell region in the target peripheral blood smear image. It is pi (π). The first in the target peripheral blood smear image i The actual area of ​​each target cell region.

[0043] It should be noted that in reality, most normal blood cells, when at rest (i.e., when observed on a blood smear), tend to appear round or nearly round. If the... i If the cells representing a target cell region are round, then... It can characterize the first i The radius of the target cell region. If the first... i If the cells representing a target cell region are round, then... It can characterize the first i The perimeter of the target cell region.

[0044] The third step is to determine the morphological irregularity of each target cell region based on the difference between the regular circular perimeter corresponding to each target cell region and the actual perimeter.

[0045] The actual perimeter of the target cell region can be represented by the total number of pixels on the edge of the target cell region.

[0046] For example, the formula for determining the morphological irregularity corresponding to the target cell region can be: ; in, The first in the target peripheral blood smear image i The morphological irregularity corresponding to each target cell region. i It is the sequence number of the target cell region in the target peripheral blood smear image. It is an absolute value function. The first in the target peripheral blood smear image i The actual perimeter of the target cell region. The first in the target peripheral blood smear image i The circumference of a regular circle corresponding to each target cell region.

[0047] It should be noted that when The larger the value, the more likely it is to indicate the first... i The more a target cell region deviates from a circular shape in its morphology, the more likely it is to exhibit swelling, deformation, or edge breakage, which often indicates a more severe degree of dissolution.

[0048] The fourth step, based on the morphological irregularity of all target cell regions in the aforementioned target peripheral blood smear image, determines the current cell distribution index, which may include the following sub-steps: The first sub-step involves dividing the target peripheral blood smear image into equal parts to obtain a reference image block.

[0049] The size of the reference image patch can be larger than the average size of all target cell regions. The reference image patch can be a rectangular region. For example, the reference image patch can be a 2mm × 2mm rectangular region.

[0050] The second sub-step involves selecting target cell regions from all target cell regions in the target peripheral blood smear image that do not intersect with the same reference image block, thus forming a set of cell regions corresponding to that reference image block.

[0051] The third sub-step is to determine the mean of the morphological irregularity of all target cell regions in the cell region set corresponding to each reference image block as the morphological irregularity representative factor for each reference image block.

[0052] The fourth sub-step involves determining the current cell distribution index based on the mean of the morphological irregularity corresponding to all target cell regions and the entropy value of the morphological irregularity representative factor corresponding to all reference image blocks.

[0053] For example, the formula for determining the current cell distribution index can be:

[0054] in,G It is an indicator of the current cell distribution. A It is the mean of the morphological irregularity of all target cell regions in the target peripheral blood smear image. H It is the entropy value of the irregularity representative factor corresponding to all reference image blocks in the target peripheral blood smear image.

[0055] It should be noted that in reality, most normal cells in the blood, when at rest (i.e., when observed on a blood smear), tend to appear round or nearly round. When A The larger the value, the more likely the target cell region in the peripheral blood smear image is to deviate from a round shape, exhibiting swelling, deformation, or fragmented edges, indicating a potentially more severe degree of dissolution. H A larger value often indicates a more random distribution of cell morphology within the target cell region of the peripheral blood smear image. This suggests a higher likelihood of severely lysed cells present in the target cell region, which are more likely to be randomly diffused spatially. Consequently, the immune system may be less able to locally control tumor cell lysis, resulting in a higher risk of microscopic loss of control. Therefore, when... G The larger the value, the more severe the dissolution of the target patient may be, and the higher the risk of the target patient evolving into tumor lysis syndrome.

[0056] The current structural disintegration potential index determination module 103 is used to determine the current structural disintegration potential index based on the gray-scale distribution and gradient distribution in the tumor CT image.

[0057] As an example, determining potential indicators of current structural collapse may include the following steps: The first step is to identify the tumor CT images of the target patient at the current time as the target tumor CT images.

[0058] The second step is to normalize the gray value corresponding to each pixel in the target tumor CT image to obtain the gray normalization factor corresponding to each pixel in the target tumor CT image.

[0059] For example, the formula for determining the gray-level normalization factor corresponding to different pixels in a target tumor CT image can be: ; in, It is the first in the target tumor CT image a The gray level normalization factor corresponding to each pixel. a It is the sequence number of the pixel in the CT image of the target tumor. It is the first in the target tumor CT image a The grayscale value corresponding to each pixel. It is the maximum gray value among all pixels in the target tumor CT image.

[0060] The third step is to normalize the gradient value corresponding to each pixel in the target tumor CT image to obtain the gradient normalization factor corresponding to each pixel in the target tumor CT image.

[0061] For example, the formula for determining the gradient normalization factor corresponding to different pixels in a target tumor CT image can be: ; in, It is the first in the target tumor CT image a The gradient normalization factor corresponding to each pixel. a It is the sequence number of the pixel in the CT image of the target tumor. It is the first in the target tumor CT image a The gradient value corresponding to each pixel. It is the maximum value among all the gradient values ​​corresponding to all pixels in the CT image of the target tumor.

[0062] The fourth step is to identify the target tumor region from the CT images of the target tumor.

[0063] For example, the U-Net network segmentation model or threshold segmentation algorithm can be used to identify tumors in CT images of target tumors, and each identified tumor region can be recorded as the target tumor region.

[0064] It should be noted that a target tumor region can characterize a single tumor. In reality, a patient may have one or more tumors in their body; therefore, there may be one or more target tumor regions.

[0065] The fifth step is to determine the possible local disintegration factors for each target tumor region based on the variance of the gray-level normalization factor corresponding to all pixels within each target tumor region and the mean of the gradient normalization factor corresponding to all pixels on the edge of each target tumor region.

[0066] For example, the formula for determining the potential factors of local disintegration in the target tumor region can be:

[0067] in, It is the first in the target tumor CT image j Local disintegration potential factors corresponding to each target tumor region. j It is the sequence number of the target tumor region in the CT image of the target tumor. In the CT images of the target tumor, the first j The variance of the gray-level normalization factor corresponding to all pixels within a target tumor region. In the CT images of the target tumor, the first j The mean of the gradient normalization factor for all pixels on the edge of a target tumor region.

[0068] It should be noted that in reality, massive lysis of tumor cells leads to necrosis and liquefaction of tumor tissue, often manifesting as uneven grayscale (interweaving of necrotic and surviving tissue) and blurred boundaries (structural disintegration) in CT images. When The larger the value, the more likely it is to indicate the first... j The more severe the interweaving of necrotic and surviving tissue within a target tumor region, the more significant the structural disintegration may be. The smaller the size, the more likely it is to indicate the first j The smaller the relative difference in grayscale at the boundaries of a target tumor region, the more blurred its boundaries may be, and the more severe the structural disintegration may be. Therefore, when The larger the value, the more likely it is to indicate the first... j The more severe the structural disintegration of a tumor in a target tumor region, the greater the release of intracellular substances (such as nucleic acids, potassium, and phosphorus), and the higher the risk of subsequent metabolic disorders.

[0069] Step 6: Based on the potential local disintegration factors corresponding to the target tumor region in the aforementioned target tumor CT images, determine the potential indicators of current structural disintegration.

[0070] For example, the cumulative value of the local disintegration potential factors corresponding to all target tumor regions in the above-mentioned target tumor CT images can be determined as the current structural disintegration potential index.

[0071] It should be noted that the higher the potential for structural disintegration, the more severe the tumor structural disintegration in the target patient is, the greater the release of intracellular substances (such as nucleic acids, potassium, and phosphorus), and the higher the risk of subsequent metabolic disorders.

[0072] The current metabolic diffusion potential indicator determination module 104 is used to determine the current metabolic diffusion potential indicator based on the changes and fluctuations of target metabolic data of different preset metabolic dimensions at all collection times during the current observation period.

[0073] As an example, determining potential indicators of current metabolic diffusion may include the following steps: The first step is to determine the representative metabolic factor corresponding to each collection time within the current observation period based on the target metabolic data of the target patient at different preset metabolic dimensions at each collection time within the current observation period.

[0074] The target metabolic data can be normalized data.

[0075] For example, any collection time within the current observation period can be designated as the marker time, and the mean of all target metabolic data of the target patient in the preset metabolic dimensions at the marker time can be designated as the metabolic representative factor corresponding to the marker time.

[0076] The second step, based on the changes and fluctuations of metabolic representative factors at each collection time and previous collection times within the current observation period, determines the metabolic instability index corresponding to each collection time within the current observation period, which may include the following sub-steps: The first sub-step involves determining the instantaneous rate of change and acceleration corresponding to each acquisition moment within the current observation period based on the metabolic representative factors corresponding to all acquisition moments within the current observation period using the central difference method.

[0077] The instantaneous rate of change is also known as the first-order central difference. Acceleration is also known as the second-order central difference. The instantaneous change at the time of data collection represents how quickly the representative metabolic factor changes at that time. The acceleration at the time of data collection represents the rate of change of the trend of change of the representative metabolic factor itself at that time.

[0078] It should be noted that the intracellular substances (uric acid, potassium, phosphorus, etc.) released by tumor cell lysis cause dynamic fluctuations in metabolic indicators, and the rate and acceleration of these changes directly reflect the urgency and worsening trend of the disorder. A higher instantaneous rate of change at the time of data collection often indicates a more likely rapid accumulation of metabolic waste and a more likely active phase of tumor lysis. A higher acceleration at the time of data collection often indicates a faster rate of deterioration of the metabolic disorder, suggesting a potential positive feedback loop of "intensified lysis - metabolic accumulation - organ dysfunction - further aggravation of metabolic accumulation," a key precursor to systemic loss of control.

[0079] The second sub-step is to determine any one of the data collection times within the current observation period as the marked time.

[0080] The third sub-step is as follows: if the instantaneous rate of change corresponding to the marked time is negative, then the target change factor corresponding to the marked time is set to 0; if the instantaneous rate of change corresponding to the marked time is not negative, then the instantaneous rate of change corresponding to the marked time is determined as the target change factor corresponding to the marked time.

[0081] The fourth sub-step is as follows: if the acceleration corresponding to the above-mentioned marked time is negative, then the reference change factor corresponding to the above-mentioned marked time is set to 0; if the acceleration corresponding to the above-mentioned marked time is not negative, then the acceleration corresponding to the above-mentioned marked time is determined as the reference change factor corresponding to the above-mentioned marked time.

[0082] The fifth sub-step involves determining the metabolic instability index corresponding to the aforementioned labeling time based on the entropy values ​​of all representative metabolic factors within the preset window period corresponding to the aforementioned labeling time, as well as the target change factor and reference change factor corresponding to the aforementioned labeling time.

[0083] The marked time can be the end time of its corresponding preset window period. The preset window period can be a pre-set time period, which can be set according to the actual scenario. For example, the preset window period can include 10 acquisition times.

[0084] For example, the formula for determining the metabolic instability index corresponding to the marker time can be: ; in, M It is an indicator of metabolic instability corresponding to the marked time. m It is the target change factor corresponding to the marked time. n It is the reference change factor corresponding to the marked time. E It is the entropy value of all metabolic representative factors within the preset window period corresponding to the marked time.

[0085] It should be noted that when metabolic regulatory mechanisms are effective, indicator fluctuations often follow physiological rhythms, with a high repetition rate of similar patterns and generally low entropy values; conversely, when regulatory mechanisms fail, indicators often exhibit irregular fluctuations, with a low repetition rate of similar patterns and increased entropy values. High entropy values, i.e. E When levels are high, it often indicates that the internal homeostasis system has lost its buffering capacity and is on the verge of decompensation, making it highly susceptible to severe metabolic disorders. m A higher value often indicates a more rapid accumulation of metabolic waste and a more likely active period for tumor lysis. n A higher value often indicates a faster rate of deterioration in metabolic disorders, suggesting a potential positive feedback loop of "increased dissolution - metabolic accumulation - organ dysfunction - further aggravation of metabolic accumulation," a key precursor to systemic loss of control. Therefore, when M The larger the value, the more severe the metabolic system may be out of control, and the closer it may be to the clinical onset of TLS.

[0086] The third step, based on the metabolic instability indicators corresponding to all collection times within the current observation period, determines the potential indicators of current metabolic diffusion, which may include the following sub-steps: The first sub-step involves dividing the current observation time period into equal parts to obtain a reference time period and a reference time period sequence.

[0087] The reference time period sequence can be a time sequence of reference time periods. The number of data collection moments within the reference time period can be preset, and can be set according to the actual scenario; for example, the number of data collection moments within the reference time period can be 3. One reference time period can represent one treatment cycle.

[0088] The second sub-step involves determining the mean value of metabolic instability indicators corresponding to all collection times within each reference time period as the representative factor of metabolic instability for each reference time period.

[0089] The third sub-step involves determining the difference between the representative metabolic instability factors corresponding to each adjacent reference time period in the above reference time period sequence as the local diffusion factor, thus obtaining the local diffusion factor sequence.

[0090] The local diffusion factor sequence can be a time series of local diffusion factors. A local diffusion factor can be equal to the representative metabolic instability factor for the next reference time period minus the representative metabolic instability factor for the previous reference time period. For example, the first local diffusion factor can be equal to the representative metabolic instability factor for the second reference time period minus the representative metabolic instability factor for the first reference time period.

[0091] It should be noted that local diffusion factors can characterize changes in representative factors of metabolic instability. The presence of positive local diffusion factors often indicates an increasing trend in representative factors of metabolic instability.

[0092] The fourth sub-step involves determining the current possible metabolic diffusion index based on the mean of all positive local diffusion factors in the aforementioned local diffusion factor sequence, the variance of the representative metabolic instability factors corresponding to all reference time periods, and the mean of the metabolic instability index corresponding to all collection times within the current observation time period.

[0093] For example, the formula for determining the current metabolic diffusion potential index can be: ; in, F It is a possible indicator of current metabolic diffusion. D It is the mean of all positive local diffusion factors in the local diffusion factor sequence; if no positive local diffusion factor exists, then we can let D It is 0. k It represents the variance of the representative factor of metabolic instability across all reference periods. Y It is the average value of metabolic instability indicators corresponding to all collection times within the current observation period.

[0094] It should be noted that when D A higher value often indicates a faster spread of metabolic disorders over time, suggesting that the metabolic products produced by dissolution are more likely to be rapidly spreading throughout the body, and that the compensatory pressure on organs is more likely to increase sharply. k A higher value often indicates more drastic fluctuations in metabolic indicators, making it less likely that the organ will be able to establish a stable compensatory mechanism, and continued fluctuations may exacerbate organ dysfunction. YA higher value often indicates a more severe degree of metabolic system dysregulation and a closer proximity to the clinical onset of TLS. Therefore, when F The larger the value, the more likely the metabolic disorder is spreading rapidly and is extremely unstable, corresponding to a high risk of serious complications such as acute kidney injury and arrhythmia in clinical practice.

[0095] The current risk stratification early warning auxiliary value determination module 105 is used to determine the current risk stratification early warning auxiliary value based on the current cell distribution index, the current structural disintegration probability index, and the current metabolic diffusion probability index.

[0096] Among them, the current risk stratification warning auxiliary value can be the risk stratification warning auxiliary value of the target patient at the current moment, which can be used to assist doctors in risk stratification and warning judgment of tumor lysis syndrome evolution.

[0097] As an example, the sum of the above-mentioned current cell distribution indicators, the above-mentioned current structural disintegration potential indicators, and the above-mentioned current metabolic diffusion potential indicators can be determined as the current risk stratification early warning auxiliary value.

[0098] It's important to note that in reality, the risk of tumor lysis syndrome (TLS) is not solely determined by microscopic loss of control or macroscopic disintegration. Microscopic loss of control often indicates a high potential for tumor lysis, while macroscopic disintegration often signifies a high intensity of tumor lysis. When the tumor structure possesses high lysis potential, it tends to further exacerbate the current tumor lysis burden. Higher current cell distribution indicators generally indicate a more severe degree of lysis in the target patient, and a higher risk of developing TLS. Higher indicators of potential structural disintegration often suggest more severe tumor structural disintegration in the target patient, potentially leading to greater release of intracellular substances (such as nucleic acids, potassium, and phosphorus), and a relatively higher risk of subsequent metabolic disorders. Current cell distribution indicators can serve as the core driving force of tumor lysis, while potential structural disintegration indicators can act as amplifying factors. When the macroscopic tumor has not yet disintegrated, the sum of the current cell distribution index and the current structural disintegration probability index is mainly driven by tumor microscopic cell lysis. When the macroscopic tumor has significantly disintegrated, the sum of the current cell distribution index and the current structural disintegration probability index often comprehensively reflects the amount and degree of tumor cell destruction and loss of control. The larger the value, the more likely the tumor lysis is highly active and out of control at both the microscopic and macroscopic levels, indicating that it has entered a high-risk pathophysiological stage of TLS (Through-Like Lesions). A larger current metabolic diffusion probability index often indicates that metabolic disorders are more likely to be spreading rapidly and are extremely unstable, corresponding to a high-risk state for serious complications such as acute kidney injury and arrhythmia in clinical practice. Therefore, a larger current risk stratification warning auxiliary value often indicates that tumor lysis may be more active, the metabolic system is more likely to be unstable and spatially diffuse, and the coupling coefficient is more likely to be significantly increased, often indicating that the system is in a high-risk state.

[0099] Optionally, risk stratification early warning auxiliary values ​​can be input as key spatiotemporal dynamic features into a pre-trained machine learning model (such as gradient boosting decision trees, deep neural networks, etc.). Through an offline training phase, the model learns the complex nonlinear mapping relationship between the risk stratification early warning auxiliary values ​​and clinical TLS incidence outcomes. During training, patient baseline physiological and clinical characteristics are simultaneously integrated as auxiliary inputs, including baseline renal function indicators, tumor type and stage, anti-cancer treatment regimens, age, and comorbidities—individual differences that enhance the model's individualized stratification adaptability. Based on the trained model, risk stratification thresholds are set using large-sample clinical follow-up data, classifying TLS risk into four levels: low, intermediate, high, and very high. The thresholds for each level can be dynamically calibrated based on the clinical data characteristics of different medical institutions. The system generates differentiated dynamic early warning signals for different risk levels, providing a scientific basis for early and precise clinical intervention and minimizing the risk of severe TLS and its complications.

[0100] refer to Figure 2Based on the same inventive concept as the above-described method embodiments, this invention provides a method for risk stratification and early warning of tumor lysis syndrome evolution, comprising the following steps: Step S1: Obtain target metabolic data of different preset metabolic dimensions of the target patient at each collection time within the current observation period, and obtain tumor CT images and peripheral blood smear images of the target patient at the current time.

[0101] Step S2: Determine the current cell distribution index based on the cell morphology in the peripheral blood smear image.

[0102] Step S3: Based on the grayscale distribution and gradient distribution in the tumor CT image, determine the possible indicators of current structural disintegration.

[0103] Step S4: Based on the changes and fluctuations of target metabolic data of different preset metabolic dimensions at all collection times during the current observation period, determine the possible indicators of current metabolic diffusion.

[0104] Step S5: Determine the current risk stratification early warning auxiliary value based on the current cell distribution index, the current structural disintegration potential index, and the current metabolic diffusion potential index.

[0105] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. For example, as shown... Figure 3 As shown, the computer device 300 includes: a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute the aforementioned method for risk stratification and early warning of tumor lysis syndrome evolution.

[0106] Based on the same inventive concept as the above-described method embodiments, the present invention provides a server, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, causing the device to execute the aforementioned method for risk stratification and early warning of tumor lysis syndrome evolution.

[0107] Based on the same inventive concept as the above-described method embodiments, the present invention provides a computer program product comprising: computer program code, which, when run on a computer, causes the computer to execute the above-described method for risk stratification and early warning of tumor lysis syndrome evolution.

[0108] Based on the same inventive concept as the above-described method embodiments, the present invention provides a computer-readable storage medium storing computer program code, which, when executed on a computer, causes the computer to perform the above-described method for risk stratification and early warning of tumor lysis syndrome evolution.

[0109] In summary, this invention quantifies current cell distribution indicators and potential structural disintegration indicators based on cell morphology in peripheral blood smear images and grayscale and gradient distribution in tumor CT images. Furthermore, by analyzing the fluctuations in various metabolic data related to the risk of tumor lysis syndrome evolution in the target patient during the current observation period, it quantifies potential metabolic diffusion indicators. This results in a quantified risk stratification warning auxiliary value to assist physicians in stratifying and predicting the risk of tumor lysis syndrome evolution, improving the rationality of the auxiliary value setting and thus enhancing the auxiliary effect.

[0110] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A risk stratification and early warning system for the evolution of tumor lysis syndrome, characterized in that, The system includes: The data acquisition module is used to acquire target metabolic data of the target patient at different preset metabolic dimensions at each acquisition time within the current observation period, and to acquire tumor CT images and peripheral blood smear images of the target patient at the current time, where the current time is the end time of the current observation period. The current cell distribution index determination module is used to determine the current cell distribution index based on the cell morphology in the peripheral blood smear image; The module for determining potential indicators of current structural disintegration is used to determine potential indicators of current structural disintegration based on the grayscale and gradient distribution in tumor CT images. The current metabolic diffusion potential indicator determination module is used to determine the current metabolic diffusion potential indicator based on the changes and fluctuations of target metabolic data of different preset metabolic dimensions at all collection times during the current observation period of the target patient. The current risk stratification early warning auxiliary value determination module is used to determine the current risk stratification early warning auxiliary value based on the current cell distribution index, the current structural disintegration probability index, and the current metabolic diffusion probability index.

2. The tumor lysis syndrome evolution risk stratification and early warning system according to claim 1, characterized in that, The process of determining the current cell distribution index based on cell morphology in peripheral blood smear images includes: The peripheral blood smear image of the target patient at the current time is identified as the target peripheral blood smear image, and cell identification is performed on the target peripheral blood smear image to obtain the target cell region; Based on the actual area of ​​each target cell region, determine the circumference of the regular circle corresponding to each target cell region; The morphological irregularity of each target cell region is determined based on the difference between the regular circular perimeter corresponding to each target cell region and the actual perimeter. The current cell distribution index is determined based on the morphological irregularity of all target cell regions in the target peripheral blood smear image.

3. The tumor lysis syndrome evolution risk stratification and early warning system according to claim 2, characterized in that, The step of determining the current cell distribution index based on the morphological irregularity of all target cell regions in the target peripheral blood smear image includes: The target peripheral blood smear image is divided into equal parts to obtain a reference image block; From all target cell regions in the target peripheral blood smear image, select target cell regions whose intersection with the same reference image block is not an empty set, and form the cell region set corresponding to the reference image block; The mean of the morphological irregularity of all target cell regions in the cell region set corresponding to each reference image block is determined as the morphological irregularity representative factor for each reference image block. The current cell distribution index is determined based on the mean of the morphological irregularity corresponding to all target cell regions and the entropy value of the morphological irregularity representative factor corresponding to all reference image patches.

4. The tumor lysis syndrome evolution risk stratification and early warning system according to claim 1, characterized in that, The determination of potential indicators of current structural disintegration based on the grayscale and gradient distribution within tumor CT images includes: The tumor CT image of the target patient at the current time is identified as the target tumor CT image; The gray value corresponding to each pixel in the target tumor CT image is normalized to obtain the gray normalization factor corresponding to each pixel in the target tumor CT image. The gradient value corresponding to each pixel in the CT image of the target tumor is normalized to obtain the gradient normalization factor corresponding to each pixel in the CT image of the target tumor. Tumor identification is performed on the CT images of the target tumor to obtain the target tumor region; Based on the variance of the gray-level normalization factor corresponding to all pixels within each target tumor region and the mean of the gradient normalization factor corresponding to all pixels on the edge of each target tumor region, determine the possible local disintegration factors for each target tumor region. Based on the potential local disintegration factors corresponding to the target tumor region in the target tumor CT image, determine the potential indicators of current structural disintegration.

5. A tumor lysis syndrome evolution risk stratification and early warning system according to claim 4, characterized in that, The step of determining the current structural disintegration potential index based on the local disintegration potential factors corresponding to the target tumor region in the target tumor CT image includes: The cumulative value of the local disintegration potential factors corresponding to all target tumor regions in the target tumor CT image is determined as the current structural disintegration potential index.

6. The tumor lysis syndrome evolution risk stratification and early warning system according to claim 1, characterized in that, The method involves determining potential indicators of metabolic diffusion based on the fluctuations in target metabolic data across different preset metabolic dimensions at all collection times within the current observation period, including: Based on the target metabolic data of the target patient at each collection time within the current observation period for different preset metabolic dimensions, determine the metabolic representative factor corresponding to each collection time within the current observation period, wherein the target metabolic data is normalized data. Based on the changes and fluctuations of metabolic representative factors at each collection time and previous collection times within the current observation period, metabolic instability indicators corresponding to each collection time within the current observation period are determined. Based on the metabolic instability indicators corresponding to all collection times within the current observation period, the possible indicators of current metabolic diffusion are determined.

7. A tumor lysis syndrome evolution risk stratification and early warning system according to claim 6, characterized in that, The step of determining the representative metabolic factor corresponding to each collection time within the current observation period based on the target metabolic data of the target patient at each collection time of different preset metabolic dimensions within the current observation period includes: Any collection time within the current observation period is designated as the marker time, and the mean of all target metabolic data of the target patient in all preset metabolic dimensions at the marker time is designated as the metabolic representative factor corresponding to the marker time.

8. A tumor lysis syndrome evolution risk stratification and early warning system according to claim 6, characterized in that, The step of determining the metabolic instability index corresponding to each collection moment in the current observation period based on the changes and fluctuations of metabolic representative factors corresponding to each collection moment and previous collection moments within the current observation period includes: Based on the metabolic representative factors corresponding to all collection times within the current observation period, the instantaneous rate of change and acceleration corresponding to each collection time within the current observation period are determined by the central difference method. Any acquisition time within the current observation period is designated as the marker time; If the instantaneous rate of change corresponding to the marked time is negative, then the target change factor corresponding to the marked time is set to 0; if the instantaneous rate of change corresponding to the marked time is not negative, then the instantaneous rate of change corresponding to the marked time is determined as the target change factor corresponding to the marked time. If the acceleration corresponding to the marked time is negative, then the reference change factor corresponding to the marked time is set to 0; if the acceleration corresponding to the marked time is not negative, then the acceleration corresponding to the marked time is determined as the reference change factor corresponding to the marked time. Based on the entropy values ​​of all representative metabolic factors within the preset window period corresponding to the marked time, as well as the target change factor and reference change factor corresponding to the marked time, the metabolic instability index corresponding to the marked time is determined, wherein the marked time is the end time of the preset window period corresponding to it.

9. A tumor lysis syndrome evolution risk stratification and early warning system according to claim 6, characterized in that, The step of determining potential metabolic diffusion indicators based on metabolic instability indicators corresponding to all collection times within the current observation period includes: Divide the current observation time period into equal parts to obtain reference time periods, and obtain a reference time period sequence; The mean value of metabolic instability indicators corresponding to all collection times within each reference period is determined as the representative factor of metabolic instability for each reference period. The difference between the representative factors of metabolic instability corresponding to each adjacent reference time period in the reference time period sequence is determined as the local diffusion factor, thus obtaining the local diffusion factor sequence; Based on the mean of all positive local diffusion factors in the local diffusion factor sequence, the variance of the representative factors of metabolic instability corresponding to all reference time periods, and the mean of metabolic instability indicators corresponding to all collection times within the current observation time period, the possible indicators of current metabolic diffusion are determined.

10. A tumor lysis syndrome evolution risk stratification and early warning system according to claim 1, characterized in that, The determination of the current risk stratification early warning auxiliary value based on the current cell distribution index, the current structural disintegration potential index, and the current metabolic diffusion potential index includes: The sum of the current cell distribution index, the current structural disintegration probability index, and the current metabolic diffusion probability index is determined as the current risk stratification early warning auxiliary value.

Images