Processing method and system for pancreatitis prognosis data

Abstract

The invention relates to a pancreatitis prognosis data processing method and system.The method comprises the steps that abdominal CT images and clinical medical examination indexes of different types of pancreatitis patients are obtained; extracting the area occupied by the seroperitoneum from the abdomen CT image; calculating the correlation between the clinical medical examination indexes and the hospitalization days of the patient; screening out clinical medical examination indexes of which the correlation is within a set range; inputting the screened clinical medical examination indexes and the area occupied by the seroperitoneum into a machine learning model for training to obtain a pancreatitis prognosis prediction model; and the pancreatitis prognosis prediction model is utilized to complete prediction of the target pancreatitis patient. According to the method, the local lesion degree and the systemic pathological state of pancreatitis can be reflected from multiple dimensions by combining abdominal CT (the area of extracted seroperitoneal effusion) and blood and urine biochemical indexes, the limitation that a traditional scoring system depends on a single data source is overcome, and the prediction precision is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and particularly to a method and system for processing pancreatitis prognosis data. Background Art

[0002] As an acute and severe disease with a high risk of death, the pathological process of pancreatitis is complex and progresses rapidly. The prognosis of patients is affected by multiple factors such as infectious necrosis, multiple organ failure, and various complications (such as pancreatic abscess, systemic inflammatory response syndrome, etc.). Clinical practice shows that early and accurate assessment of the patient's prognosis plays a decisive role in formulating individualized treatment plans (such as the timing of necrotic tissue debridement surgery and the dynamic allocation of intensive care resources), which can significantly reduce the mortality rate and improve the clinical outcome.

[0003] Currently, the prognosis assessment of pancreatitis mainly relies on traditional scoring systems, such as the Ranson score, the APACHE-II (Acute Physiology and Chronic Health Evaluation II), and the BISAP score, etc. These systems calculate the weighted values through clinically selected indicators (such as age, laboratory parameters, imaging features), and then stratify and predict the patient's risk of death or the probability of complication occurrence. However, such methods have significant limitations: 1. Insufficient timeliness: Traditional scoring systems rely on static indicators 24 - 48 hours after admission and cannot track the dynamic changes of the condition in real time, resulting in the assessment results lagging behind the actual pathological process.

[0004] 2. Low prediction accuracy: The linear weighted model is difficult to capture the non-linear associations between multiple factors, resulting in limited prediction accuracy. Summary of the Invention

[0005] To solve the above problems, the purpose of the embodiments of the present invention is to provide a method and system for processing pancreatitis prognosis data.

[0006] A method for processing pancreatitis prognosis data includes: Step 1: Obtain the abdominal CT images and clinical medical test indicators of patients with different types of pancreatitis; the clinical medical test indicators include: blood routine, liver and kidney function electrolytes, blood amylase, urine amylase, erythrocyte sedimentation rate, and C-reactive protein; Step 2: Extract the area occupied by abdominal effusion from the abdominal CT images; Step 3: Calculate the correlation between the clinical medical test indicators and the patient's length of hospital stay; Step 4: Screen out the clinical medical test indicators with the correlation within a set range; Step 5: Input the screened clinical medical test indicators and the area occupied by abdominal effusion into a machine learning model for training to obtain a pancreatitis prognosis prediction model; Step 6: Use the pancreatitis prognosis prediction model to complete the prediction of the target pancreatitis patient.

[0007] Preferably, the step 2: Extract the area occupied by peritoneal effusion from the abdominal CT image, including: Step 2.1: Take a four-neighborhood and an eight-neighborhood centered on any point on the abdominal CT image; Step 2.2: Calculate the gradient means on the four-neighborhood and the eight-neighborhood respectively; Step 2.3: Take the gradient mean on the four-neighborhood as the first element and the gradient mean on the eight-neighborhood as the second element to form a binary tuple; Step 2.4: Construct an edge detection model with the binary tuple as the object, and use the edge detection model to complete the detection of the abdominal CT image; Step 2.5: Extract the edge occupied by peritoneal effusion; Step 2.6: Determine the area occupied by peritoneal effusion based on the number of pixel points within the edge occupied by peritoneal effusion.

[0008] Preferably, the step 2.4: Construct an edge detection model with the binary tuple as the object, and use the edge detection model to complete the detection of the abdominal CT image, including: Step 2.4.1: Construct an initial threshold, and divide the binary tuple to obtain a first interval formula and a second interval formula; Step 2.4.2: Quantify the gradient difference of the abdominal CT image based on the first interval formula and the second interval formula to form a first interval difference formula and a second interval difference formula; Step 2.4.3: Construct an edge detection model based on the first interval difference formula and the second interval difference formula; Step 2.4.4: Solve the maximum value of the objective function to make the difference between the first interval and the second interval the largest to obtain the image edge division threshold; Step 2.4.5: Use the image edge division threshold to complete the detection of the abdominal CT image.

[0009] Preferably, the step 2.4.1: Construct an initial threshold, and divide the binary tuple to obtain a first interval formula and a second interval formula, including: Solve the number of occurrences of each binary tuple, and use the initial threshold to divide the number of occurrences of each binary tuple to obtain a first interval formula and a second interval formula; where, the first interval formula and the second interval formula are:

[0010] Wherein, represents the number of occurrences of each binary tuple, represents a binary tuple, whenp When = 1, it represents the average gradient of the four-neighborhood corresponding to the first point on the abdominal CT image. When q When = 1, it represents the average gradient of the eight-neighborhood corresponding to the first point on the abdominal CT image. represents the initial threshold. s represents the average gradient of the four-neighborhood corresponding to the sth point on the abdominal CT image, and k represents the average gradient of the eight-neighborhood corresponding to the kth point on the abdominal CT image. represents the length of the image. represents the width of the image. represents the first interval formula. represents the second interval formula. represents the density of the occurrence of the binary tuple. L is the gray level.

[0011] Preferably, in the step 2.4.2, the first interval difference formula and the second interval difference formula are:

[0012]

[0013] Among them, represents the first interval difference formula. represents the second interval difference formula. represents a binary tuple. When p When = 1, it represents the average gradient of the four-neighborhood corresponding to the first point on the abdominal CT image. When q When = 1, it represents the average gradient of the eight-neighborhood corresponding to the first point on the abdominal CT image. represents the density of the occurrence of the binary tuple.

[0014] Preferably, in the step 2.4.3: constructing an edge detection model based on the first interval difference formula and the second interval difference formula, including: Using the mean square error to quantify the gradient difference between the first interval difference formula and the second interval difference formula to obtain the edge detection model; among them, the edge detection model is:

[0015] Among them, represents the edge detection model. represents the overall mean of the first interval. represents the overall mean of the second interval.

[0016] Preferably, in the step 3: calculating the correlation between the clinical medical test indicators and the patient's length of hospital stay, including: Step 3.1: Determine whether each clinical medical test index and the patient's length of hospital stay satisfy the normal distribution; Step 3.2: Calculate the correlation between the clinical medical test indexes that satisfy the normal distribution and the patient's length of hospital stay using the Pearson correlation coefficient; Step 3.3: Calculate the correlation between the clinical medical test indexes that do not satisfy the normal distribution and the patient's length of hospital stay using the rank coefficient; where the rank coefficient calculation formula is:

[0017] where, represents the rank difference between the two variables, represents the total number of samples, represents the correlation between the clinical medical test index and the patient's length of hospital stay.

[0018] The present invention also provides a processing system for pancreatitis prognosis data, including: A data acquisition module, configured to acquire abdominal CT images and clinical medical test indexes of different types of pancreatitis patients; the clinical medical test indexes include: blood routine, liver and kidney function electrolytes, blood amylase, urine amylase, erythrocyte sedimentation rate, and C-reactive protein; An area extraction module, configured to extract the area occupied by abdominal effusion from the abdominal CT image; A correlation calculation module, configured to calculate the correlation between the clinical medical test indexes and the patient's length of hospital stay; An index screening module, configured to screen out the clinical medical test indexes whose correlation is within a set range; A training module, configured to input the screened clinical medical test indexes and the area occupied by abdominal effusion into a machine learning model for training to obtain a pancreatitis prognosis prediction model; A prediction module, configured to complete the prediction of the target pancreatitis patient using the pancreatitis prognosis prediction model.

[0019] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are connected through the bus. It is characterized in that when the computer program is executed by the processor, the steps in the above-mentioned processing method for pancreatitis prognosis data are implemented.

[0020] The present invention also provides a computer-readable storage medium, on which a computer program is stored. It is characterized in that when the computer program is executed by a processor, the steps in the above-mentioned processing method for pancreatitis prognosis data are implemented.

[0021] According to the specific embodiments provided by the present invention, the following technical effects are disclosed by the present invention: The present invention relates to a method and system for processing pancreatitis prognosis data. Compared with the prior art, by combining abdominal CT (extracting the area of peritoneal effusion) and blood and urine biochemical indexes, the present invention can reflect the local lesion degree and systemic pathological state of pancreatitis from multiple dimensions, overcome the limitation of the traditional scoring system relying on a single data source, and significantly improve the prediction accuracy.

[0022] To make the above objects, features and advantages of the present invention more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, makes a detailed description as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0024] Figure 1 It is a flowchart of a method for processing pancreatitis prognosis data provided by the present invention; Figure 2 It is a schematic diagram of peritoneal effusion and length of hospital stay provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0026] In the present invention, unless otherwise clearly defined and limited, the terms "installed", "connected", "connected", "fixed", etc. should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0027] Please refer to Figure 1 , a method for processing pancreatitis prognosis data, including: Step 1: Obtain the abdominal CT images and clinical medical examination indicators of patients with different types of pancreatitis; the clinical medical examination indicators include: blood routine, liver and kidney function electrolytes, blood amylase, urine amylase, erythrocyte sedimentation rate, and C-reactive protein; Step 2: Extract the area occupied by peritoneal effusion from the abdominal CT images; As Figure 2 shown, peritoneal effusion, as one of the complications of pancreatitis, contains a large number of toxic substances such as cytokines, vasoactive substances, and pancreatic enzymes. The chain reaction caused by these toxic substances is one of the main reasons for early multiple organ dysfunction syndrome and death in patients. It is of great significance for early and rapid identification of the severity of their condition and the risk of death.

[0028] Further, the said Step 2 includes: Step 2.1: Take a four-neighborhood and an eight-neighborhood with any point on the abdominal CT image as the center; Step 2.2: Calculate the gradient means on the four-neighborhood and the eight-neighborhood respectively; Step 2.3: Take the gradient mean on the four-neighborhood as the first element and the gradient mean on the eight-neighborhood as the second element to form a binary group; Step 2.4: Construct an edge detection model with the binary group as the object, and use the edge detection model to complete the detection of the abdominal CT image; In Step 2.4, it includes: Step 2.4.1: Construct an initial threshold, and divide the binary group to obtain a first interval formula and a second interval formula; In the said Step 2.4.1, solve the number of occurrences of each binary group, and use the initial threshold to divide the number of occurrences of each binary group to obtain a first interval formula and a second interval formula; where, the first interval formula and the second interval formula are:

[0029] where, represents the number of occurrences of each binary group, represents the initial threshold, represents the length of the image, represents the width of the image, represents the first interval formula, represents the second interval formula, represents the density of the occurrence of the binary group, L is the gray level.

[0030] Step 2.4.2: Quantify the gradient difference of the abdominal CT image based on the first interval formula and the second interval formula to form the first interval difference formula and the second interval difference formula; where, the first interval difference formula and the second interval difference formula are:

[0031]

[0032] Where, represents the first interval difference formula, represents the second interval difference formula, represents a binary tuple. When p = 1, it represents the gradient mean of the four-neighborhood corresponding to the first point on the abdominal CT image. When q = 1, it represents the gradient mean of the eight-neighborhood corresponding to the first point on the abdominal CT image. represents the initial threshold, s represents the gradient mean of the four-neighborhood corresponding to the sth point on the abdominal CT image, and k represents the gradient mean of the eight-neighborhood corresponding to the kth point on the abdominal CT image.

[0033] The four-neighborhood usually refers to the four adjacent pixels above, below, left, and right of a pixel, while the eight-neighborhood includes the four pixels of the four-neighborhood plus the four pixels in the diagonal directions. When calculating the gradient, the gradient of the four-neighborhood may pay more attention to the horizontal and vertical edges, while the eight-neighborhood can capture more edges in different directions because it contains diagonal direction information. The present invention combines the gradients of both to comprehensively utilize the gradient information in multiple directions and improve the sensitivity to edges in different directions.

[0034] Step 2.4.3: Construct an edge detection model based on the first interval difference formula and the second interval difference formula; In the step 2.4.3, the mean square error is used to quantify the gradient difference between the first interval difference formula and the second interval difference formula to obtain the edge detection model; where, the edge detection model is:

[0035] Where, represents the edge detection model, represents the overall mean of the first interval, represents the overall mean of the second interval.

[0036] Step 2.4.4: Solve the maximum value of the objective function to make the difference between the first interval and the second interval the largest to obtain the image edge division threshold; Step 2.4.5: Use the image edge division threshold to complete the detection of the abdominal CT image.

[0037] The maximum value of the gradient difference in the present invention reflects the significant changes between the edge regions. By combining multi-neighborhood information (four-neighborhood + eight-neighborhood), the problem of edge loss caused by insignificant gradients in a single direction can be reduced.

[0038] Step 2.5: Extract the edges occupied by the peritoneal effusion; Step 2.6: Determine the area occupied by the peritoneal effusion based on the number of pixel points within the edges occupied by the peritoneal effusion.

[0039] Step 3: Calculate the correlation between the clinical medical examination indicators and the patient's length of hospital stay; Furthermore, the said Step 3 includes: Step 3.1: Determine whether each clinical medical examination indicator and the patient's length of hospital stay satisfy the normal distribution; Step 3.2: Use the Pearson correlation coefficient to calculate the correlation between the clinical medical examination indicators that satisfy the normal distribution and the patient's length of hospital stay; Step 3.3: Use the rank coefficient to calculate the correlation between the clinical medical examination indicators that do not satisfy the normal distribution and the patient's length of hospital stay; wherein, the formula for the rank coefficient is:

[0040] Wherein, represents the rank difference between two variables, represents the total number of samples, represents the correlation between the clinical medical examination indicator and the patient's length of hospital stay.

[0041] Step 4: Screen out the clinical medical examination indicators whose correlation is within the set range; Step 5: Input the screened clinical medical examination indicators and the area occupied by the peritoneal effusion into the machine learning model for training to obtain a pancreatitis prognosis prediction model; Step 6: Use the pancreatitis prognosis prediction model to complete the prediction of the target pancreatitis patient.

[0042] By combining abdominal CT (extracting the area of peritoneal effusion) and blood and urine biochemical indicators, the present invention can reflect the local lesion degree and systemic pathological state of pancreatitis from multiple dimensions, overcome the limitation of the traditional scoring system relying on a single data source, and significantly improve the prediction accuracy.

[0043] The present invention also provides a processing system for pancreatitis prognosis data, including: A data acquisition module, used to acquire abdominal CT images and clinical medical examination indicators of different types of pancreatitis patients; the clinical medical examination indicators include: blood routine, liver and kidney function electrolytes, blood amylase, urine amylase, erythrocyte sedimentation rate, and C-reactive protein; An area extraction module for extracting the area occupied by peritoneal effusion from abdominal CT images; A correlation calculation module for calculating the correlation between clinical medical test indicators and the length of hospital stay of patients; An index screening module for screening out clinical medical test indicators with a correlation within a set range; A training module for inputting the screened clinical medical test indicators and the area occupied by peritoneal effusion into a machine learning model for training to obtain a pancreatitis prognosis prediction model; A prediction module for using the pancreatitis prognosis prediction model to complete the prediction of target pancreatitis patients.

[0044] An embodiment of the present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected through the bus. When the computer program is executed by the processor, it realizes each process in the embodiment of the above-mentioned method for processing pancreatitis prognosis data and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0045] In addition, an embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, it realizes each process in the embodiment of the above-mentioned method for processing pancreatitis prognosis data and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0046] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technical solution that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for processing pancreatitis prognosis data, characterized in that: include: Step 1: Obtain abdominal CT images and clinical medical examination indicators of patients with different types of pancreatitis; The clinical medical examination indicators include: blood routine, liver and kidney function electrolytes, blood amylase, urine amylase, erythrocyte sedimentation rate and C-reactive protein; Step 2: Extract the area occupied by the peritoneal effusion from the abdominal CT image; Step 3: Calculate the correlation between clinical medical examination indicators and the number of days of hospitalization of patients; Step 4: Screen out clinical medical test indicators with correlation within the set range; Step 5: Input the screened clinical medical test indicators and the area occupied by the peritoneal effusion into the machine learning model for training to obtain a pancreatitis prognosis prediction model; Step 6: Use the pancreatitis prognosis prediction model to complete the prediction of target pancreatitis patients.

2. A method for processing pancreatitis prognosis data according to claim 1, characterized in that: The step 2: extracting the area occupied by the peritoneal effusion from the abdominal CT image, comprises: Step 2.1: Take a four-neighborhood and an eight-neighborhood with any point on the abdominal CT image as the center; Step 2.2: Calculate the mean gradient of the four-neighborhood and eight-neighborhood respectively; Step 2.3: Take the mean gradient of the four neighborhoods as the first element and the mean gradient of the eight neighborhoods as the second element to form a tuple; Step 2.4: constructing an edge detection model with the binary group as the object, and using the edge detection model to complete the detection of the abdominal CT image; Step 2.5: Extract the margin occupied by the peritoneal effusion; Step 2.6: Determine the area occupied by the peritoneal effusion based on the number of pixels within the edge occupied by the peritoneal effusion.

3. A method for processing pancreatitis prognosis data according to claim 2, characterized in that: The step 2.4: constructing an edge detection model with the binary as the object, and using the edge detection model to complete the detection of the abdominal CT image, includes: Step 2.4.1: Construct an initial threshold value and divide the binary group into the first interval formula and the second interval formula; Step 2.4.2: quantifying the gradient difference of the abdominal CT image based on the first interval formula and the second interval formula to form a first interval difference formula and a second interval difference formula; Step 2.4.3: Construct an edge detection model based on the first interval difference formula and the second interval difference formula; Step 2.4.4: Solve the maximum value of the objective function to maximize the difference between the first interval and the second interval to obtain the image edge division threshold; Step 2.4.5: Use the image edge division threshold to complete the detection of the abdominal CT image.

4. A method for processing pancreatitis prognosis data according to claim 3, characterized in that: The step 2.4.1: constructing an initial threshold value, dividing the binary group into the first interval formula and the second interval formula, includes: Solve the number of occurrences of each bigram, and use the initial threshold to divide the number of occurrences of each bigram to obtain the first interval formula and the second interval formula; the first interval formula and the second interval formula are: ； in, represents the number of occurrences of each bigram. represents the initial threshold, Represents a two-tuple, when p =1 indicates the mean gradient of the four neighborhoods corresponding to the first point on the abdominal CT image. q =1 indicates the mean gradient of the eight neighborhoods corresponding to the first point on the abdominal CT image. represents the initial threshold, s represents the mean gradient of the four neighborhoods corresponding to the s-th point on the abdominal CT image, k represents the mean gradient of the eight neighborhoods corresponding to the k-th point on the abdominal CT image, Indicates the length of the image, Indicates the width of the image. represents the first interval formula, represents the second interval formula, represents the density of occurrence of two-tuples, L is grayscale.

5. A method for processing pancreatitis prognosis data according to claim 4, characterized in that: In step 2.4.2, the first interval difference formula and the second interval difference formula are: ； ； in, represents the first interval difference formula, represents the second interval difference formula, Represents a two-tuple, when p =1 indicates the mean gradient of the four neighborhoods corresponding to the first point on the abdominal CT image. q =1 indicates the mean gradient of the eight neighborhoods corresponding to the first point on the abdominal CT image. Represents the density of occurrence of bigrams.

6. A method for processing pancreatitis prognosis data according to claim 5, characterized in that: The step 2.4.3: constructing an edge detection model based on the first interval difference formula and the second interval difference formula, comprises: The edge detection model is obtained by quantifying the gradient difference between the first interval difference formula and the second interval difference formula using the mean square error; wherein the edge detection model is: ； in, represents the edge detection model, represents the population mean of the first interval, represents the population mean of the second interval.

7. A method for processing pancreatitis prognosis data according to claim 6, characterized in that: The step 3: calculating the correlation between clinical medical examination indicators and the number of days of hospitalization of patients, including: Step 3.1: Determine whether each clinical medical test index and the number of days of hospitalization of the patient satisfy the normal distribution; Step 3.2: Use the Pearson correlation coefficient to calculate the correlation between clinical medical test indicators that meet the normal distribution and the patient's hospitalization days; Step 3.3: Use the grade coefficient to calculate the correlation between the clinical medical test indicators that meet the non-normal distribution and the patient's hospitalization days; the grade coefficient calculation formula is: ； in, represents the rank difference between two variables, represents the total number of samples, Represents the correlation between clinical medical examination indicators and the number of days of hospitalization of patients.

8. A system for processing pancreatitis prognosis data, characterized in that: include: A data acquisition module is used to obtain abdominal CT images and clinical medical examination indicators of patients with different types of pancreatitis; The clinical medical examination indicators include: blood routine, liver and kidney function electrolytes, blood amylase, urine amylase, erythrocyte sedimentation rate and C-reactive protein; An area extraction module is used to extract the area occupied by the peritoneal effusion from the abdominal CT image; A correlation calculation module is used to calculate the correlation between clinical medical test indicators and the number of hospitalization days of patients; The indicator screening module is used to screen out clinical medical test indicators whose correlation is within the set range; A training module is used to input the screened clinical medical test indicators and the area occupied by the peritoneal effusion into the machine learning model for training to obtain a pancreatitis prognosis prediction model; The prediction module is used to complete the prediction of target pancreatitis patients using the pancreatitis prognosis prediction model.

9. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps in the method for processing pancreatitis prognosis data according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method for processing pancreatitis prognosis data according to any one of claims 1 to 7 are implemented.

Images