Placenta implantation risk assessment system based on data analysis

Abstract

The invention relates to the technical field of placenta implantation risk assessment, and particularly discloses a placenta implantation risk assessment system based on data analysis, which comprises a data integration module, a data extraction module, a data analysis module and a data screening module, and is characterized in that placenta implantation data of a to-be-delivered maternal body is extracted based on image data of the to-be-delivered maternal body; the preoperative examination data of the to-be-delivered maternal body is combined to predict the bleeding amount of the to-be-delivered maternal body in the operation; therefore, the safety and the management efficiency in the operation can be improved, the improvement of the doctor-patient relationship and the progress of the medical technology can be promoted, and the system is an indispensable important component in modern obstetrical management, can help doctors to formulate personalized operation schemes in advance, and can improve the operation efficiency of the maternal to be delivered. Comprising the steps of estimation of blood reserve, configuration of an operation team and preparation of emergency response measures, so that the perinatal death rate and the complication occurrence rate of a maternal body are remarkably reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of placenta accreta risk assessment and relates to a placenta accreta risk assessment system based on data analysis. Background Art

[0002] Abnormal placenta accreta (such as placenta accreta, placenta accreta, or placenta accreta disease) is one of the main causes of massive obstetric hemorrhage. Massive intraoperative bleeding not only directly threatens maternal life but can also lead to serious complications such as hemorrhagic shock, disseminated intravascular coagulation (DIC), and even multi-organ failure. Therefore, the importance of accurately assessing placenta accreta before delivery and predicting the potential amount of intraoperative bleeding based on preoperative examination data is highlighted.

[0003] Existing technologies have multiple deficiencies in placenta accreta risk assessment, primarily manifested in incomplete data analysis, limited prediction accuracy, and insufficient individualization of risk assessment. First, traditional placenta accreta risk assessment often relies on a single imaging examination or limited clinical indicators. This approach cannot fully capture the complex biological interactions between the placenta and uterus, resulting in potentially inaccurate assessment results. Second, existing technologies typically rely on empirical rules or simple statistical models to predict intraoperative blood loss. This approach is inadequate when dealing with multiple variables and nonlinear relationships, making it difficult to accurately predict blood loss. Summary of the Invention

[0004] In response to the aforementioned challenges in the existing technologies, the present invention provides a data-analysis-based placenta accreta risk assessment system. This system aims to predict intraoperative blood loss and assess risk levels by collecting and analyzing maternal placenta accreta information. By combining medical imaging analysis, patient clinical data, and machine learning algorithms, the present invention achieves a precise assessment of placenta accreta risk, addressing the aforementioned technical issues.

[0005] In order to achieve the above-mentioned and other purposes, the technical solutions adopted by the present invention are as follows: In one aspect, the present invention provides a placenta accreta risk assessment system based on data analysis, comprising a data integration module, a data extraction module, a data analysis module, and a data screening module. The aforementioned modules are connected via wired and / or wireless connections to achieve data transmission between the modules. Data integration module: collects imaging data of the expectant mother and preoperative examination data; Data extraction module: extracting the placenta implantation data of the expectant mother based on the image data of the expectant mother; Data analysis module: Based on the placenta implantation data of the expectant mother and the preoperative examination data of the expectant mother, the intraoperative bleeding volume of the expectant mother is predicted; Data screening module: Based on the predicted value of the intraoperative bleeding volume of the expectant mother, the surgical risk level of the expectant mother is graded and assessed.

[0006] The placenta accreta data of the expectant mother include the thickness of the placenta, the depth of adhesion of the placenta to the uterine wall, and the area of ​​placenta invading the uterine myometrium.

[0007] The thickness of the placenta corresponding to the mother to be born, the specific collection process is as follows: The image data of the expectant mother is converted into a standardized grayscale image, and the corresponding standardized grayscale image of the expectant mother is preprocessed. The specific preprocessing operations are divided into image enhancement and noise reduction. The specific processing process is as follows: Adaptive histogram equalization is used to enhance the contrast of the placenta region. The formula is: ,in is the original pixel value of the image data corresponding to the mother to be born, is the pixel value of the enhanced image data corresponding to the mother to be born, are the standard deviation and mean of the global image respectively; are the mean and standard deviation of the local window respectively; The non-local mean filter is used to remove the noise of the corresponding standardized grayscale image of the mother to be born. The formula is: ;in is the grayscale value of pixel y, is the weight between pixels, which is defined based on similarity: , are pixel blocks centered at x and y respectively; h is the filtering parameter, Represents a pixel block centered at x and y and The square of the Euclidean distance between them; represents the pixel value of the image at position x after non-local mean filtering, e is a natural constant, and Ω represents the pixel range of the image; A U-Net-based deep learning segmentation model was used to extract the preliminary region of the placenta. An edge detection algorithm was then used to extract the inner and outer boundaries of the placenta region. The estimated thickness T(x,y) of the placenta corresponding to the expectant mother is defined as the shortest distance between the inner and outer boundaries of the placenta: , is the Euclidean distance, is the set of points on the outer boundary of the placenta, where (x, y) is the coordinate of any point on the inner boundary of the placenta, and (x', y') is the coordinate of any point on the outer boundary of the placenta; Finally, the estimated thickness T(x,y) of the placenta corresponding to the expectant mother is summed up and the mean is calculated to obtain the average thickness of the placenta corresponding to the expectant mother. ; At the same time, considering the influence of the image tilt angle on the thickness calculation, the correction formula , the thickness T of the placenta corresponding to the mother to be delivered is obtained, and θ is the angle between the imaging plane and the vertical direction of the placenta.

[0008] The depth of adhesion between the placenta and the uterine wall of the mother to be born. The specific collection process is as follows: Use 3D U-Net to segment the placenta and uterine wall, and then output the placenta region probability map and uterine wall area probability map ; Extract the regions of the placenta and uterine wall based on the probability threshold: ; ; Where R1 and R2 represent the three-dimensional spatial regions of the placenta and uterine wall, respectively; YZ1 and YZ2 represent the set placenta probability threshold and uterine wall probability threshold, respectively; Then, the Sobel operator is used to extract the boundaries of the placenta and uterine wall from the areas of the placenta and uterine wall; The point of adhesion between the placenta and the uterine wall is defined as the point on the inner boundary of the placenta B1 with the smallest distance from the outer boundary of the uterine wall B2: , NL is the adhesion point between the placenta and the uterine wall, p and q are any points on the boundary between the placenta and the uterine wall, respectively; The adhesion depth D(x,y,z) is defined as the shortest distance from a point within the placental region to the border of the uterine wall: , represents any point on the boundary of the uterine wall; Then calculate the average adhesion depth D, the specific formula is as follows: , is the weight function within the placenta region; At the same time, considering the effect of angle deviation on adhesion depth during image acquisition and the elastic deformation of placenta and uterine wall tissue, the correction formula Calculate the adhesion depth between the placenta and the uterine wall of the mother to be born , θ' is the angle between the imaging plane and the vertical direction of the placenta and uterine wall, and △D is the depth correction value set to avoid elastic deformation.

[0009] The area of ​​placenta invading the myometrium of the expectant mother is collected as follows: Use 3D U-Net to segment the placenta and uterine wall, and then output the placenta region probability map and myometrial region probability map ; Then the three-dimensional spatial area of ​​the placenta and uterine myometrium is obtained; The area where the placenta invades the myometrium is defined as the overlapping area of ​​the three-dimensional space of the placenta and the three-dimensional space of the myometrium: R4=R1∩R3, R3 is the three-dimensional spatial area of ​​the myometrium; R4 is the area where the placenta invades the myometrium; Then, the Monte Carlo method is used to calculate the preliminary area of ​​placenta invading myometrium of the expectant mother based on the area of ​​placenta invading myometrium. At the same time, considering the influence of the image acquisition process on the area and the elastic deformation of the placenta and uterine myometrium, the correction formula Calculate the area of ​​placenta invading the uterine myometrium of the expectant mother , θ'' is the angle between the imaging plane and the plane of the placenta and uterine myometrium, △A is the area correction value set to avoid elastic deformation, and A is the initial area of ​​the placenta invading the uterine myometrium of the mother to be delivered.

[0010] Predict the amount of bleeding during the operation of the expectant mother. The specific prediction process is as follows: The preoperative examination data of the expectant mother include placental blood flow velocity, blood flow resistance index and placental vascular distribution density; Importing placenta accreta data from expectant mothers into a blood loss prediction model The predicted value Q of the amount of bleeding during the operation of the expectant mother is obtained. in They represent the direct effect of the area of ​​placenta invading the myometrium on the amount of intraoperative blood loss. , α is the adjustment factor for the prediction of placental thickness and bleeding volume, , α0 is the basic proportional coefficient, ω is the thickness adjustment factor, reflecting the marginal effect of placental thickness; λ1 is the placenta type factor, adjusted according to the implantation type; They represent the direct effect of the depth of adhesion between the placenta and the uterine wall on the amount of intraoperative blood loss. , e is a natural constant, λ2 is the depth influence coefficient, which controls the amplitude of the influence of depth on bleeding volume, λ3 is the depth attenuation coefficient, which reflects the nonlinear effect of depth on bleeding volume, and λ4 is the regional distribution factor, which reflects the shape and distribution characteristics of the depth of placental adhesion; They represent the direct effect of the area of ​​placenta invading the myometrium on the amount of intraoperative blood loss. , λ5 is the vascular density influence coefficient, which controls the influence of vascular density on bleeding volume, BD is the placental vascular distribution density of the expectant mother, λ6 is the vascular distribution heterogeneity factor, which reflects the distribution characteristics of blood vessels in the placenta, RI is the blood flow resistance index of the expectant mother, and λ7 is the resistance regulation coefficient, which controls the regulatory effect of blood flow resistance on bleeding volume; They represent the direct effect of placental blood flow velocity on intraoperative blood loss, , λ8 is the blood flow velocity influence coefficient, which controls the influence of blood flow velocity on bleeding volume, υ is the blood flow velocity adjustment factor, which is used to adjust the differences between individuals, V is the placental blood flow velocity of the mother to be delivered, and η is the velocity smoothing factor, which prevents the formula from failing when the blood flow velocity is zero.

[0011] The surgical risk level of the expectant mother is assessed in a graded manner. The specific assessment process is as follows: Based on the predicted value of the amount of bleeding during surgery for the expectant mother, it is compared with the predicted value interval of bleeding volume corresponding to each surgical risk level. If the predicted value of the amount of bleeding during surgery for the expectant mother is within the predicted value interval of bleeding volume corresponding to a certain surgical risk level, then the surgical risk level will be used as the surgical risk level of the expectant mother.

[0012] Another aspect of the present invention provides a placenta accreta risk assessment device based on data analysis, comprising a processor, a memory, and a communication bus; The memory stores a computer-readable program executable by the processor; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, the processor implements the placenta accreta risk assessment system based on data analysis according to the present invention.

[0013] As described above, the placenta accreta risk assessment system based on data analysis provided by the present invention has at least the following beneficial effects: The data-analysis-based placenta accreta risk assessment system provided by this invention extracts placenta accreta-related data from maternal imaging data before delivery and, in combination with preoperative examination data, predicts intraoperative blood loss, thereby grading the surgical risk level. This method has important clinical significance and necessity in modern obstetrics. First, abnormal placenta accreta (such as placenta accreta, placenta accreta, or placenta accreta disease) is one of the main causes of obstetric hemorrhage. Intraoperative hemorrhage not only directly threatens maternal life but can also cause serious complications such as hemorrhagic shock, disseminated intravascular coagulation (DIC), and even multiple organ failure. Therefore, accurately assessing placenta accreta before delivery and predicting the expected intraoperative blood loss based on preoperative examination data can help physicians formulate personalized surgical plans in advance, including blood reserve estimation, surgical team deployment, and emergency response preparation, thereby significantly reducing maternal perinatal mortality and complication rates.

[0014] Combined with preoperative examination data, this allows for a more comprehensive assessment of the mother's physiological status and potential intraoperative challenges. By integrating imaging data with preoperative examination data to develop a predictive model, intraoperative blood loss can be more accurately estimated. This predictive capability not only aids preoperative preparation (such as blood collection and detailed surgical planning), but also provides crucial insights for anesthesia and intraoperative monitoring, ensuring the surgical team can rapidly respond to emergencies and reduce the occurrence of intraoperative complications. It also facilitates personalized medical management. For example, for high-risk patients, experienced surgical and anesthesia teams can be deployed in advance, along with more comprehensive intraoperative monitoring equipment and emergency response measures. For low-risk patients, the process can be streamlined to reduce unnecessary resource waste. This tiered management approach improves the efficiency of medical resource utilization, reduces healthcare costs, and enhances patient safety and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 Schematic diagram of the connection of various modules of the system of the present invention. DETAILED DESCRIPTION

[0017] The above contents described below in conjunction with the implementation of the present invention are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they shall fall within the scope of protection of the present invention.

[0018] Example 1: Please refer to Figure 1 As shown, the placenta accreta risk assessment system based on data analysis includes a data integration module, a data extraction module, a data analysis module and a data screening module. The above modules are connected by wired and / or wireless connections to achieve data transmission between the modules; Data integration module: collects imaging data of the expectant mother and preoperative examination data; Data extraction module: extracting the placenta implantation data of the expectant mother based on the image data of the expectant mother; The placenta accreta data of the expectant mother include the thickness of the placenta, the depth of adhesion of the placenta to the uterine wall, and the area of ​​placenta invading the uterine myometrium.

[0019] The thickness of the placenta corresponding to the mother to be born, the specific collection process is as follows: The image data of the expectant mother is converted into a standardized grayscale image, and the corresponding standardized grayscale image of the expectant mother is preprocessed. The specific preprocessing operations are divided into image enhancement and noise reduction. The specific processing process is as follows: Adaptive histogram equalization is used to enhance the contrast of the placenta region. The formula is: ,in is the original pixel value of the image data corresponding to the mother to be born, is the pixel value of the enhanced image data corresponding to the mother to be born, are the standard deviation and mean of the global image respectively; are the mean and standard deviation of the local window respectively; The non-local mean filter is used to remove the noise of the corresponding standardized grayscale image of the mother to be born. The formula is: ;in is the grayscale value of pixel y, is the weight between pixels, which is defined based on similarity: , are pixel blocks centered at x and y respectively. Indicates the contribution of y to x; h is a filtering parameter used to control the sensitivity of similarity; Represents a pixel block centered at x and y and The square of the Euclidean distance between them; represents the pixel value at position x of the image after non-local mean filtering, e is a natural constant, Ω represents the pixel range of the image, usually a search window centered on x; In the calculation formula, if two pixel blocks are very similar, the weight will be close to 1; if they are not similar, the weight will be close to 0. This weight is used to weight the average pixel value during the filtering process to achieve the denoising effect; The denominator in the above calculation formula is used to normalize the weights to ensure that the sum of all weights is 1; The goal of the non-local mean filter (NLM) is to calculate smoother and more denoised pixel values ​​by considering the similarity of all pixels in the image. Unlike traditional local filtering (such as mean filtering and Gaussian filtering), NLM relies not only on the spatial distance between pixels but also considers the content similarity of pixel blocks. For each pixel x, NLM takes a weighted average of the values ​​I(y) of the surrounding pixels based on its similarity weight ω(x, y) with other pixels to obtain a smoothed pixel value. The above calculation formula achieves denoising by calculating the weighted average of all pixels in the original image (or pixels within a certain range), where the weights are determined by the similarity of the pixel blocks. This method can better preserve image details while reducing noise, making it particularly suitable for denoising natural and medical images.

[0020] A U-Net-based deep learning segmentation model was used to extract the preliminary region of the placenta. An edge detection algorithm was then used to extract the inner and outer boundaries of the placenta region. The estimated thickness T(x,y) of the placenta corresponding to the expectant mother is defined as the shortest distance between the inner and outer boundaries of the placenta: , is the Euclidean distance, is the set of points on the outer boundary of the placenta, where (x, y) is the coordinate of any point on the inner boundary of the placenta, and (x', y') is the coordinate of any point on the outer boundary of the placenta; It should be added that the inner boundary of the placenta usually refers to the boundary close to the fetus (such as the junction of the placenta and the amniotic cavity), and the outer boundary of the placenta usually refers to the boundary close to the uterus (such as the junction of the placenta and the uterine wall).

[0021] Finally, the estimated thickness T(x,y) of the placenta corresponding to the expectant mother is summed up and the mean is calculated to obtain the average thickness of the placenta corresponding to the expectant mother. ; At the same time, considering the influence of the image tilt angle on the thickness calculation, the correction formula , the thickness T of the placenta corresponding to the mother to be delivered is obtained, and θ is the angle between the imaging plane and the vertical direction of the placenta.

[0022] The depth of adhesion between the placenta and the uterine wall of the mother to be born is collected as follows: Use 3D U-Net to segment the placenta and uterine wall, and then output the placenta region probability map and uterine wall area probability map ; Extract the regions of the placenta and uterine wall based on the probability threshold: ; ; Where R1 and R2 represent the three-dimensional spatial regions of the placenta and uterine wall, respectively; YZ1 and YZ2 represent the set placenta probability threshold and uterine wall probability threshold, respectively; YZ1 and YZ2 represent the probability thresholds used for segmentation, respectively. Only when the probability is greater than YZ1 and YZ2, the point will be considered to belong to the placenta region and the uterine wall region; Then, the Sobel operator is used to extract the boundaries of the placenta and uterine wall from the areas of the placenta and uterine wall; The point of adhesion between the placenta and the uterine wall is defined as the point on the inner boundary of the placenta B1 with the smallest distance from the outer boundary of the uterine wall B2: , NL is the adhesion point between the placenta and the uterine wall, p and q are any points on the boundary between the placenta and the uterine wall, respectively; The adhesion depth D(x,y,z) is defined as the shortest distance from a point within the placental region to the border of the uterine wall: , represents any point on the boundary of the uterine wall; Then calculate the average adhesion depth D, the specific formula is as follows: , is the weight function within the placenta region; At the same time, considering the effect of angle deviation on adhesion depth during image acquisition and the elastic deformation of placenta and uterine wall tissue, the correction formula Calculate the adhesion depth between the placenta and the uterine wall of the mother to be born , θ' is the angle between the imaging plane and the vertical direction of the placenta and uterine wall, and △D is the depth correction value set to avoid elastic deformation.

[0023] The area of ​​placenta invading the myometrium of the expectant mother is collected as follows: Use 3D U-Net to segment the placenta and uterine wall, and then output the placenta region probability map and myometrial region probability map ; Then the three-dimensional spatial area of ​​the placenta and uterine myometrium is obtained; The area where the placenta invades the myometrium is defined as the overlapping area of ​​the three-dimensional space of the placenta and the three-dimensional space of the myometrium: R4=R1∩R3, R3 is the three-dimensional spatial area of ​​the myometrium; R4 is the area where the placenta invades the myometrium; Then, the Monte Carlo method is used to calculate the preliminary area of ​​placenta invading myometrium of the expectant mother based on the area of ​​placenta invading myometrium. At the same time, considering the influence of the image acquisition process on the area and the elastic deformation of the placenta and uterine myometrium, the correction formula Calculate the area of ​​placenta invading the uterine myometrium of the expectant mother , θ'' is the angle between the imaging plane and the plane of the placenta and uterine myometrium, △A is the area correction value set to avoid elastic deformation, and A is the initial area of ​​the placenta invading the uterine myometrium of the mother to be delivered.

[0024] Data analysis module: Based on the placenta implantation data of the expectant mother and the preoperative examination data of the expectant mother, the intraoperative bleeding volume of the expectant mother is predicted; Predict the amount of bleeding during the operation of the expectant mother. The specific prediction process is as follows: The preoperative examination data of the expectant mother include placental blood flow velocity, blood flow resistance index and placental vascular distribution density; Importing placenta accreta data from expectant mothers into a blood loss prediction model The predicted value Q of the amount of bleeding during the operation of the expectant mother is obtained. in They represent the direct effect of the area of ​​placenta invading the myometrium on the amount of intraoperative blood loss. , α is the adjustment factor for the prediction of placental thickness and bleeding volume, , α0 is the basic proportional coefficient, ω is the thickness adjustment factor, reflecting the marginal effect of placental thickness; λ1 is the placenta type factor, adjusted according to the implantation type; In practical terms, the larger the area of ​​placenta accreta, the wider the area that needs to be dissected during surgery, the more blood vessels are damaged, and therefore the amount of bleeding will increase. Different types of placenta accreta (such as partial accreta and complete accreta) will affect the amount of bleeding in this area.

[0025] They represent the direct effect of the depth of adhesion between the placenta and the uterine wall on the amount of intraoperative blood loss. , e is a natural constant, λ2 is the depth influence coefficient, which controls the amplitude of the influence of depth on bleeding volume, λ3 is the depth attenuation coefficient, which reflects the nonlinear effect of depth on bleeding volume, and λ4 is the regional distribution factor, which reflects the shape and distribution characteristics of the depth of placental adhesion; In practical terms, the greater the depth of adhesion, the more severe the placenta accreta, the greater the number and depth of blood vessels damaged during placental removal, and the greater the amount of bleeding. The effect of adhesion depth is nonlinear; after a certain depth, its marginal effect on bleeding volume decreases.

[0026] They represent the direct effect of the area of ​​placenta invading the myometrium on the amount of intraoperative blood loss. , λ5 is the vascular density influence coefficient, which controls the influence of vascular density on bleeding volume, BD is the placental vascular distribution density of the expectant mother, λ6 is the vascular distribution heterogeneity factor, which reflects the distribution characteristics of blood vessels in the placenta, RI is the blood flow resistance index of the expectant mother, and λ7 is the resistance regulation coefficient, which controls the regulatory effect of blood flow resistance on bleeding volume; In practical terms, the denser the blood vessels in the placenta, the greater the amount of bleeding when the vessels are damaged during surgery. The blood flow resistance index (RI) modifies this effect, as higher blood flow resistance may reduce bleeding after vascular rupture.

[0027] They represent the direct effect of placental blood flow velocity on intraoperative blood loss, , λ8 is the blood flow velocity influence coefficient, which controls the influence of blood flow velocity on bleeding volume, υ is the blood flow velocity adjustment factor, which is used to adjust the differences between individuals, V is the placental blood flow velocity of the mother to be delivered, and η is the velocity smoothing factor, which prevents the formula from failing when the blood flow velocity is zero; In practical terms, the higher the placental blood flow rate, the faster the blood will flow out after a blood vessel ruptures, leading to increased bleeding. However, when the blood flow rate reaches a certain level, the increase in bleeding volume will tend to slow down.

[0028] Data screening module: Based on the predicted value of the intraoperative bleeding volume of the expectant mother, the surgical risk level of the expectant mother is graded and assessed.

[0029] The surgical risk level of the expectant mother is assessed in a graded manner. The specific assessment process is as follows: Based on the predicted value of the amount of bleeding during surgery for the expectant mother, it is compared with the predicted value interval of bleeding volume corresponding to each surgical risk level. If the predicted value of the amount of bleeding during surgery for the expectant mother is within the predicted value interval of bleeding volume corresponding to a certain surgical risk level, then the surgical risk level will be used as the surgical risk level of the expectant mother.

[0030] Example 2: A placenta accreta risk assessment device based on data analysis, comprising a processor, a memory, and a communication bus; The memory stores a computer-readable program executable by the processor; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, the processor implements the placenta accreta risk assessment system based on data analysis according to the present invention.

[0031] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0032] It should be understood that determining B based on A does not mean determining B based solely on A. B can also be determined based on A and / or other information.

[0033] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0034] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A placenta accreta risk assessment system based on data analysis, characterized in that: It includes a data integration module, a data extraction module, a data analysis module and a data screening module. The above modules are connected by wired and / or wireless connections to achieve data transmission between the modules; Data integration module: collects imaging data of the expectant mother and preoperative examination data; Data extraction module: extracting the placenta implantation data of the expectant mother based on the image data of the expectant mother; Data analysis module: Based on the placenta implantation data of the expectant mother and the preoperative examination data of the expectant mother, the intraoperative bleeding volume of the expectant mother is predicted; Data screening module: Based on the predicted value of the intraoperative bleeding volume of the expectant mother, the surgical risk level of the expectant mother is graded and assessed.

2. The placenta accreta risk assessment system based on data analysis according to claim 1, characterized in that: The placenta accreta data of the expectant mother include the thickness of the placenta, the depth of adhesion of the placenta to the uterine wall, and the area of ​​placenta invading the uterine myometrium.

3. The placenta accreta risk assessment system based on data analysis according to claim 2, characterized in that: The thickness of the placenta corresponding to the mother to be born, the specific collection process is as follows: The image data of the expectant mother is converted into a standardized grayscale image, and the corresponding standardized grayscale image of the expectant mother is preprocessed. The specific preprocessing operations are divided into image enhancement and noise reduction. The specific processing process is as follows: Adaptive histogram equalization is used to enhance the contrast of the placenta region. The formula is: ,in is the original pixel value of the image data corresponding to the mother to be born, is the pixel value of the enhanced image data corresponding to the mother to be born, are the standard deviation and mean of the global image respectively; are the mean and standard deviation of the local window respectively; The non-local mean filter is used to remove the noise of the corresponding standardized grayscale image of the mother to be born. The formula is: ;in is the grayscale value of pixel y, is the weight between pixels, which is defined based on similarity: , are pixel blocks centered at x and y respectively; h is the filtering parameter, Represents a pixel block centered at x and y and The square of the Euclidean distance between them; represents the pixel value of the image at position x after non-local mean filtering, e is a natural constant, and Ω represents the pixel range of the image; A U-Net-based deep learning segmentation model was used to extract the preliminary region of the placenta. An edge detection algorithm was then used to extract the inner and outer boundaries of the placenta region. The estimated thickness T(x,y) of the placenta corresponding to the expectant mother is defined as the shortest distance between the inner and outer boundaries of the placenta: , is the Euclidean distance, is the set of points on the outer boundary of the placenta, where (x, y) is the coordinate of any point on the inner boundary of the placenta, and (x', y') is the coordinate of any point on the outer boundary of the placenta; Finally, the estimated thickness T(x,y) of the placenta corresponding to the expectant mother is summed up and the mean is calculated to obtain the average thickness of the placenta corresponding to the expectant mother. ; At the same time, considering the influence of the image tilt angle on the thickness calculation, the correction formula , the thickness T of the placenta corresponding to the mother to be delivered is obtained, and θ is the angle between the imaging plane and the vertical direction of the placenta.

4. The placenta accreta risk assessment system based on data analysis according to claim 2, characterized in that: The depth of adhesion between the placenta and the uterine wall of the mother to be born is collected as follows: Use 3D U-Net to segment the placenta and uterine wall, and then output the placenta region probability map and uterine wall area probability map ; Extract the regions of the placenta and uterine wall based on the probability threshold: ； ； Where R1 and R2 represent the three-dimensional spatial regions of the placenta and uterine wall, respectively; YZ1 and YZ2 represent the set placenta probability threshold and uterine wall probability threshold, respectively; Then, the Sobel operator is used to extract the boundaries of the placenta and uterine wall from the areas of the placenta and uterine wall; The point of adhesion between the placenta and the uterine wall is defined as the point on the inner boundary of the placenta B1 with the smallest distance from the outer boundary of the uterine wall B2: , NL is the adhesion point between the placenta and the uterine wall, p and q are any points on the boundary between the placenta and the uterine wall, respectively; The adhesion depth D(x,y,z) is defined as the shortest distance from a point within the placental region to the border of the uterine wall: , represents any point on the boundary of the uterine wall; Then calculate the average adhesion depth D, the specific formula is as follows: , is the weight function within the placenta region; At the same time, considering the effect of angle deviation on adhesion depth during image acquisition and the elastic deformation of placenta and uterine wall tissue, the correction formula Calculate the adhesion depth between the placenta and the uterine wall of the mother to be born , θ' is the angle between the imaging plane and the vertical direction of the placenta and uterine wall, and △D is the depth correction value set to avoid elastic deformation.

5. The placenta accreta risk assessment system based on data analysis according to claim 2, characterized in that: The area of ​​placenta invading the myometrium of the expectant mother is collected as follows: Use 3D U-Net to segment the placenta and uterine wall, and then output the placenta region probability map and myometrial region probability map ; Then the three-dimensional spatial area of ​​the placenta and uterine myometrium is obtained; The area where the placenta invades the myometrium is defined as the overlapping area of ​​the three-dimensional space of the placenta and the three-dimensional space of the myometrium: R4=R1∩R3, R3 is the three-dimensional spatial area of ​​the myometrium; R4 is the area where the placenta invades the myometrium; Then, the Monte Carlo method is used to calculate the preliminary area of ​​placenta invading myometrium of the expectant mother based on the area of ​​placenta invading myometrium. Taking into account the influence of the image acquisition process on the area and the elastic deformation of the placenta and uterine myometrium, the correction formula Calculate the area of ​​placenta invading the uterine myometrium of the expectant mother , θ'' is the angle between the imaging plane and the plane of the placenta and uterine myometrium, △A is the area correction value set to avoid elastic deformation, and A is the initial area of ​​the placenta invading the uterine myometrium of the mother to be delivered.

6. The placenta accreta risk assessment system based on data analysis according to claim 1, characterized in that: Predict the amount of bleeding during the operation of the expectant mother. The specific prediction process is as follows: The preoperative examination data of the expectant mother include placental blood flow velocity, blood flow resistance index and placental vascular distribution density; Importing placenta accreta data from expectant mothers into a blood loss prediction model The predicted value Q of the amount of bleeding during the operation of the expectant mother is obtained. in They represent the direct effect of the area of ​​placenta invading the myometrium on the amount of intraoperative blood loss. , α is the adjustment factor for the prediction of placental thickness and bleeding volume, , α0 is the basic proportional coefficient, ω is the thickness adjustment factor, reflecting the marginal effect of placental thickness; λ1 is the placenta type factor, adjusted according to the implantation type; They represent the direct effect of the depth of adhesion between the placenta and the uterine wall on the amount of intraoperative blood loss. , e is a natural constant, λ2 is the depth influence coefficient, which controls the amplitude of the influence of depth on bleeding volume, λ3 is the depth attenuation coefficient, which reflects the nonlinear effect of depth on bleeding volume, and λ4 is the regional distribution factor, which reflects the shape and distribution characteristics of the depth of placental adhesion; They represent the direct effect of the area of ​​placenta invading the myometrium on the amount of intraoperative blood loss. , λ5 is the vascular density influence coefficient, which controls the influence of vascular density on bleeding volume, BD is the placental vascular distribution density of the expectant mother, λ6 is the vascular distribution heterogeneity factor, which reflects the distribution characteristics of blood vessels in the placenta, RI is the blood flow resistance index of the expectant mother, and λ7 is the resistance regulation coefficient, which controls the regulatory effect of blood flow resistance on bleeding volume; They represent the direct effect of placental blood flow velocity on intraoperative blood loss, , λ8 is the blood flow velocity influence coefficient, which controls the influence of blood flow velocity on bleeding volume, υ is the blood flow velocity adjustment factor, which is used to adjust the differences between individuals, V is the placental blood flow velocity of the mother to be delivered, and η is the velocity smoothing factor, which prevents the formula from failing when the blood flow velocity is zero.

7. The placenta accreta risk assessment system based on data analysis according to claim 1, characterized in that: The surgical risk level of the expectant mother is assessed in a graded manner. The specific assessment process is as follows: Based on the predicted value of the amount of bleeding during surgery for the expectant mother, it is compared with the predicted value interval of bleeding volume corresponding to each surgical risk level. If the predicted value of the amount of bleeding during surgery for the expectant mother is within the predicted value interval of bleeding volume corresponding to a certain surgical risk level, then the surgical risk level will be used as the surgical risk level of the expectant mother.

8. A placenta accreta risk assessment device based on data analysis, characterized in that: It is implemented based on the placenta accreta risk assessment system based on data analysis according to any one of claims 1 to 7, and includes a processor, a memory and a communication bus; The memory stores a computer-readable program executable by the processor; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, it is implemented to implement the placenta accreta risk assessment system based on data analysis according to any one of claims 1 to 7.

Images