A preoperative anesthesia risk prediction method, device and equipment for PPGL and a storage medium

Abstract

This application discloses a PPGL preoperative anesthesia risk prediction method, device, equipment, and storage medium, relating to the field of preoperative anesthesia risk prediction. The method includes: acquiring multi-dimensional perioperative data; determining a first risk score based on demographic and medical history data; determining a second risk score based on physiological sign time-series data; determining a third risk score based on laboratory test data; determining a fourth risk score based on preoperative intervention and medication data; and determining a fifth risk score based on perioperative planning data; determining a preoperative risk score based on the first, second, third, fourth, and fifth risk scores; and obtaining a prediction result indicating no preoperative anesthesia risk if the preoperative risk score is greater than or equal to a risk threshold. This method can reduce the risk of serious perioperative cardiovascular events and protect patient safety.

Description

Technical Field

[0001] This application relates to the field of preoperative anesthesia risk prediction technology, and in particular to a PPGL preoperative anesthesia risk prediction method, device, equipment and storage medium. Background Technology

[0002] Pheochromocytoma and paraganglioma (PPGL) are rare but highly dangerous neuroendocrine tumors, and their perioperative management is extremely challenging. Patients often experience severe hemodynamic fluctuations and endocrine metabolic disorders. If preoperative risk assessment is inadequate, surgery can easily trigger serious cardiovascular events, endangering the patient's life.

[0003] Currently, existing technologies for preoperative risk prediction in PPGL patients largely rely on subjective judgment based on the long-term accumulated clinical experience of anesthesiologists, or employ a single-dimensional scoring system constructed based on a fixed number of clinical indicators. These approaches attempt to identify high-risk groups through simple quantitative standards.

[0004] However, existing solutions generally suffer from inaccurate predictions, an inability to dynamically adapt to individual differences, and a lack of attribution guidance. Summary of the Invention

[0005] This application provides a method for predicting preoperative anesthesia risk in PPGL surgery, which can dynamically adapt to individual differences and provides attribution guidance, thereby improving the accuracy of predicting preoperative anesthesia risk in PPGL surgery.

[0006] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides a PPGL preoperative anesthesia risk prediction method, including: Obtain perioperative multidimensional data for the i-th patient. The perioperative multidimensional data includes: demographic and medical history data, physiological signs time sequence data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. Based on demographic and medical history data, the first risk score of the i-th patient is determined; based on physiological signs and time series data, the second risk score of the i-th patient is determined; based on laboratory test data, the third risk score of the i-th patient is determined; based on preoperative intervention and medication data, the fourth risk score of the i-th patient is determined; and based on perioperative planning data, the fifth risk score of the i-th patient is determined. Based on the first risk score, second risk score, third risk score, fourth risk score and fifth risk score of the i-th patient, determine the preoperative risk score of the i-th patient; If the preoperative risk score of the i-th patient is greater than or equal to the risk threshold of the i-th patient, then the prediction result is that the i-th patient has no preoperative anesthesia risk.

[0007] Optionally, the method further includes: If the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient, then the prediction result of the i-th patient having preoperative anesthesia risk is obtained.

[0008] Optionally, the method further includes: Obtain the clinical core score of the d-th dimension, the average confidence score of all fields in the d-th dimension for the i-th patient, and the average population information gain score of all fields in the d-th dimension for the i-th patient, and determine the weight of the d-th dimension for the i-th patient. The preoperative risk score for the i-th patient is determined based on the first, second, third, fourth, and fifth risk scores, including: Using the weights of the i-th patient in each dimension, the first risk score, second risk score, third risk score, fourth risk score, and fifth risk score of the i-th patient are weighted and summed to obtain the preoperative risk score of the i-th patient.

[0009] Optionally, the risk threshold for the i-th patient is determined in the following manner: Obtain the baseline risk threshold, the average confidence level of all fields, and the baseline risk factor for the i-th patient; The risk threshold for patient i is determined based on the baseline risk threshold for patient i, the average confidence level of all fields, and the baseline risk factor.

[0010] Optionally, the method further includes: If the preoperative risk score of patient i is less than the risk threshold of patient i, calculate the contribution of each dimension of patient i to the overall risk and the contribution of each field in each dimension to the overall risk. A risk assessment report is generated based on the contribution of each dimension of the i-th patient to the overall risk and the contribution of each field in each dimension to the overall risk.

[0011] Optionally, obtaining the perioperative multidimensional data of the i-th patient includes: Obtain the initial multidimensional data for the i-th patient; The initial multidimensional data of the i-th patient are standardized to obtain the perioperative multidimensional data of the i-th patient.

[0012] Optionally, the method further includes: Store the perioperative multidimensional data, preoperative risk score, and prediction results for the i-th patient.

[0013] Secondly, this application provides a PPGL preoperative anesthesia risk prediction device, comprising: The multi-dimensional data acquisition module is used to acquire perioperative multi-dimensional data for the i-th patient. The perioperative multi-dimensional data includes: demographic and medical history data, physiological signs time series data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. The multidimensional risk scoring module is used to determine the first risk score of the i-th patient based on demographic and medical history data, the second risk score of the i-th patient based on physiological signs time sequence data, the third risk score of the i-th patient based on laboratory test data, the fourth risk score of the i-th patient based on preoperative intervention and medication data, and the fifth risk score of the i-th patient based on perioperative planning data. The preoperative risk score determination module is used to determine the preoperative risk score of the i-th patient based on the first risk score, second risk score, third risk score, fourth risk score, and fifth risk score of the i-th patient. The prediction result determination module is used to obtain a prediction result that the i-th patient has no preoperative anesthesia risk when the preoperative risk score of the i-th patient is greater than or equal to the risk threshold of the i-th patient.

[0014] Thirdly, this application provides a computing device, including a memory and a processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of the first aspects.

[0015] Fourthly, this application provides a computer-readable storage medium for storing a computer program for performing the method as described in any one of the first aspects.

[0016] As can be seen from the above technical solution, this application has at least the following beneficial effects: In this application, perioperative multi-dimensional data of the i-th patient is obtained. This data includes: demographic and medical history data, physiological sign time-series data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. Based on the demographic and medical history data, a first risk score is determined for the i-th patient; based on the physiological sign time-series data, a second risk score is determined; based on the laboratory test data, a third risk score is determined; based on the preoperative intervention and medication data, a fourth risk score is determined; and based on the perioperative planning data, a fifth risk score is determined. Based on the first, second, third, fourth, and fifth risk scores, a preoperative risk score for the i-th patient is determined. If the preoperative risk score for the i-th patient is greater than or equal to the patient's risk threshold, a prediction result indicating that the i-th patient has no preoperative anesthesia risk is obtained.

[0017] This invention effectively addresses the problems of inaccurate predictions, difficulty in adapting to individual differences, and lack of attribution guidance caused by existing technologies that rely on physicians' subjective experience and single-dimensional scoring systems. This application collects data from five dimensions: demographics and medical history, physiological signs and their time sequence, laboratory tests, preoperative intervention and medication, and perioperative planning. Corresponding risk scores are calculated for each dimension and then integrated to obtain an individualized preoperative risk score. This score is then combined with patient-specific risk thresholds for risk assessment, achieving a shift from subjective experience-based judgment to objective quantitative evaluation. The multi-dimensional data collection and stratified scoring model comprehensively cover the key risk factors for PPGL patients, improving the comprehensiveness and accuracy of preoperative anesthesia risk assessment for PPGL, while dynamically adapting to individual differences among patients. This reduces the risk of serious perioperative cardiovascular events, ensuring patient safety, and establishes a standardized assessment process, which is beneficial for clinical promotion and anesthesia quality control.

[0018] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0019] Figure 1 A flowchart of a PPGL preoperative anesthesia risk prediction method provided in this application embodiment; Figure 2 A schematic diagram of a PPGL preoperative anesthesia risk prediction device provided in an embodiment of this application; Figure 3 This is a schematic diagram of a computing device provided in an embodiment of this application. Detailed Implementation

[0020] The terms "first," "second," and "third," etc., used in this application specification and accompanying drawings are used to distinguish different objects, not to limit a specific order.

[0021] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0022] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the related technologies is given first: PPGL is an abbreviation for Pheochromocytoma and Paraganglioma, a type of neuroendocrine tumor that secretes catecholamines. Preoperative anesthesia management is challenging, and hemodynamic fluctuations pose a significant risk. Preoperative anesthesia risk prediction is a technical process that quantifies potential adverse events such as drastic blood pressure fluctuations, arrhythmias, cardiovascular accidents, and shock during anesthesia, based on clinical data, laboratory indicators, and medical history. PPGL preoperative anesthesia risk prediction is a specialized, quantitative, and traceable clinical assessment technique for perioperative safety in PPGL patients. It is used to identify high-risk individuals, pinpoint risk factors, and guide preoperative optimization and anesthesia protocol development before anesthesia administration.

[0023] In current technologies, preoperative risk prediction for PPGL patients largely relies on the subjective judgment of anesthesiologists based on their long-term accumulated clinical experience, or uses a single-dimensional scoring system constructed based on a fixed number of clinical indicators. These approaches attempt to identify high-risk groups through simple quantitative standards. However, such approaches generally suffer from inaccurate prediction results, an inability to dynamically adapt to individual differences, and a lack of attribution guidance.

[0024] In view of this, embodiments of this application provide a PPGL preoperative anesthesia risk prediction method, which can be executed by a processing device. This processing device can be a terminal or a server. Terminals include, but are not limited to, smartphones, tablets, laptops, personal digital assistants, or smart wearable devices. The server can be a cloud server, such as a central server in a central cloud computing cluster or an edge server in an edge cloud computing cluster. Alternatively, the server can be a server in a local data center. A local data center refers to a data center directly controlled by the user.

[0025] This method is based on the pathophysiological characteristics of PPGL, takes multi-dimensional clinical information as input, takes standardized, dynamic risk scores and dynamic thresholds as the core of calculation, and takes the prediction results of preoperative anesthesia risk as output, thus constructing a disease-specific preoperative risk prediction method that is traceable throughout the process and adaptive to individuals. Specifically, this application provides a PPGL preoperative anesthesia risk prediction method. This method first acquires perioperative multi-dimensional data for the i-th patient, including: demographic and medical history data, physiological sign time-series data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. Then, based on the demographic and medical history data, a first risk score is determined for the i-th patient; based on the physiological sign time-series data, a second risk score is determined; based on the laboratory test data, a third risk score is determined; based on the preoperative intervention and medication data, a fourth risk score is determined; and based on the perioperative planning data, a fifth risk score is determined. Finally, based on the first, second, third, fourth, and fifth risk scores, the preoperative risk score for the i-th patient is determined. If the preoperative risk score for the i-th patient is greater than or equal to the patient's risk threshold, a prediction result indicating that the i-th patient has no preoperative anesthesia risk is obtained.

[0026] To make the technical solution of this application clearer and easier to understand, the following is combined with... Figure 1 This application introduces a PPGL preoperative anesthesia risk prediction method provided in its embodiments. For example... Figure 1 As shown, this figure is a flowchart of a PPGL preoperative anesthesia risk prediction method provided in an embodiment of this application. The method includes: S201, Obtain perioperative multidimensional data for the i-th patient.

[0027] The perioperative multidimensional data includes: demographic and medical history data, physiological signs and their time sequence data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. Specifically, the demographic and medical history data includes: age, sex, height, weight, ASA classification, previous surgical history, previous anesthesia history, history of comorbidities, and PPGL. Data includes pathological diagnosis information, tumor localization information, etc.; physiological signs time-series data includes: daily measurements of systolic blood pressure, diastolic blood pressure, heart rate, and pulse oxygen saturation within 7 days before surgery, as well as corresponding measurement time and method; laboratory test data includes: catecholamine and metabolite test values, complete blood count, blood biochemistry, coagulation function, electrolytes, liver and kidney function, and myocardial marker test values ​​within 30 days before surgery, as well as corresponding test time, test method, and reference range; preoperative intervention and medication data includes: types, dosages, durations, and medication adjustment records of antihypertensive drugs, alpha-blockers, beta-blockers, and volume expansion therapy within 30 days before surgery; perioperative planning data includes: planned surgical procedure, planned anesthesia method, estimated surgical duration, estimated intraoperative blood loss range, and surgical team qualification level information.

[0028] It should be noted that the acquisition of multi-dimensional perioperative data from patients must strictly comply with relevant laws and regulations in the medical and health industry. For any sensitive personal information of patients, separate, explicit, and written informed consent has been obtained from the patients before acquisition, and the data is limited to data analysis and assessment related to PPGL preoperative anesthesia risk prediction, and will not be used beyond the authorized scope. Data processing adheres to the principles of legality, legitimacy, necessity, and integrity, implementing de-identification, anonymization, and data desensitization processes, and adopting security protection measures such as encrypted storage, access control, and operation logging to ensure that data is not leaked, tampered with, destroyed, or illegally provided or used.

[0029] It should be noted that, in the embodiments provided in this application, the perioperative multidimensional data of the i-th patient can be obtained by acquiring the initial multidimensional data of the i-th patient and then standardizing the initial multidimensional data of the i-th patient to obtain the perioperative multidimensional data of the i-th patient.

[0030] The process of obtaining the initial multidimensional data of the i-th patient can be as follows: perform semantic mapping between field names and standard field libraries on the collected raw data of each dimension to generate a raw dataset with unified field identifiers; wherein, the standard field library is pre-built based on the medical standard terminology system, and each standard field corresponds to a unique field code, data type, and value range.

[0031] The process of standardizing the initial multidimensional data of the i-th patient can be as follows: perform data standardization and confidence quantification on the generated original dataset with uniform field identifiers, which is the initial multidimensional data of the i-th patient, and output a standardized dataset. This standardized dataset can be used as a representation of the perioperative multidimensional data of the i-th patient.

[0032] Specifically, the standardization process for numerical data is as follows: For all numerical fields, unit standardization is performed first, followed by numerical normalization to generate standardized values. Unit standardization involves converting all values ​​to their corresponding standard units based on a pre-built unit conversion table for different units of measurement corresponding to the same standard field. Numerical normalization uses a min-max normalization method to process the standardized values; the specific calculation formula is as follows:

[0033] in, The standardized value of the j-th standard field for the i-th patient; This represents the original value of the j-th standard field for the i-th patient after unit standardization. The minimum value of the physiological value range corresponding to the predefined j-th standard field; This represents the maximum value within the physiological value range corresponding to the predefined j-th standard field. If... Exceeding [ , The interval is truncated according to the interval boundary value, and outlier values ​​are marked at the same time.

[0034] The standardization process for time-series data is as follows: First, convert all time fields to standard timestamps in YYYYMMDD format. For example, January 1, 2000, after conversion to a standard timestamp is 20000101, preserving the original time information. Second, using the surgery date as the baseline node, calculate the preoperative time interval corresponding to each time-series data point. The unit is days; third, for physiological signs time series data, extract time series statistical features, including mean, range, coefficient of variation, and fluctuation frequency, and generate standardized values ​​of time series features.

[0035] The standardization process for categorized data is as follows: for categorized fields, based on a pre-built classification coding mapping table, the classification values ​​of text types are converted into ordered coding values ​​to generate standardized classification codes.

[0036] The field confidence metric standardization process is as follows: For each standardized field, calculate the corresponding confidence value using the following formula:

[0037] in, Let be the confidence value of the j-th standard field for the i-th patient, with a value range of [0,1]. , , For the predefined confidence weight coefficients, satisfying ; The confidence level for the data source is predefined based on the system type of the data source, and its value range is [0,1]. The confidence score for time validity is calculated based on the interval between the field detection time and the surgery date, using the following formula: k is a predefined time decay coefficient. The number of days between the detection time and the surgery date corresponding to the j-th field; To standardize compliance confidence, the value is 1000 if the field value does not exceed the physiological value range and is not missing; the value is 0 if there are missing values; and the value is 0.5 if there are outliers that are truncated.

[0038] To address the shortcomings of existing technologies in predicting factors and dimensions that are limited, this application collects data covering five dimensions: demographics and medical history, physiological signs and time series, laboratory tests, preoperative intervention and medication, and perioperative planning. This comprehensive approach covers various risk factors related to the pathophysiological characteristics of PPGL and perioperative management, avoiding biases in risk prediction results caused by missing dimensions.

[0039] S202, based on perioperative multidimensional data, determines multiple risk scores for the i-th patient.

[0040] Specifically, based on demographic and medical history data, the first risk score of the i-th patient is determined; based on physiological signs and time-series data, the second risk score of the i-th patient is determined; based on laboratory test data, the third risk score of the i-th patient is determined; based on preoperative intervention and medication data, the fourth risk score of the i-th patient is determined; and based on perioperative planning data, the fifth risk score of the i-th patient is determined.

[0041] In the embodiments provided in this application, as an optional implementation method, when determining multiple risk scores for the i-th patient based on perioperative multidimensional data, the weight of each field in each dimension of data can be determined first, and then the risk score corresponding to the dimension of data can be determined according to the fields contained in each dimension and their corresponding weights.

[0042] For each standard field j, its dynamic weight The calculation formula is:

[0043] Where M is the total number of standard fields within the dimension to which the field belongs; α, β, γ, and δ are predefined hyperparameters, where α is the weighting coefficient of field information gain, β is the weighting coefficient of field clinical relevance, γ is the weighting coefficient of field confidence, and δ is the time adaptation decay coefficient, used to control the degree of influence of detection time deviation on weight, satisfying α+β+γ=1, δ>0. Let be the population information gain value of the j-th standard field, pre-calculated using the information gain algorithm based on the PPGL patient preoperative risk event history dataset. Its value ranges from [0,1] and is used to characterize the distinguishing ability of this field for risk events. Specifically, it can be calculated using the following formula:

[0044] in, For the original information gain of the j-th field, through get, The overall entropy of the historical dataset; The conditional entropy is given when the j-th field is known. The minimum value among the original information gains of all standard fields; The maximum value among the original information gains of all standard fields; Assign a value to the clinical correlation of the j-th standard field. Based on the pathophysiological characteristics of PPGL and the predefined clinical guidelines, the value range is [0,1], which is used to characterize the pathophysiological correlation strength between this field and the preoperative risk of PPGL. This represents the confidence value of the j-th standard field for the i-th patient. This represents the number of days between the detection time and the surgery date corresponding to the j-th field; The optimal preoperative testing interval days recommended by clinical guidelines for the j-th field are predefined in the standard field library; This is a time-adaptation factor used to dynamically adjust field weights based on the deviation between the detection time and the optimal detection interval. The larger the deviation, the smaller the factor value, and the lower the corresponding field weight. For the actual time interval of the m-th field, The optimal time interval recommended by clinical guidelines for the m-th field; m is the traversal index of the fields within the dimension, used to sum all fields within the same dimension; This represents the group information gain value of the m-th field within the dimension. Assign a value to the clinical relevance of the m-th field within the dimension; For the i-th patient, this is the field confidence value of the m-th field within the dimension.

[0045] It should be noted that, in the embodiments provided in this application, the mechanism by which each dimension affects the risk score may be as follows: For the demographic and medical history dimension, the fields of this dimension reflect the patient's baseline physical condition, exposure to comorbidities, and the baseline anesthesia risk corresponding to the ASA classification, thus affecting the calculation of risk factor values; the more comorbidities and the higher the ASA classification, the higher the risk factor value of the corresponding field, and the lower the dimension risk score, indicating a higher risk associated with this dimension; for the physiological signs time-series dimension, the fields of this dimension reflect the patient's preoperative hemodynamic fluctuations and the degree of abnormality in baseline vital signs, thus affecting the calculation of risk factor values; the greater the blood pressure fluctuation and the higher the degree of heart rate abnormality, the higher the risk factor value of the corresponding field, and the lower the dimension risk score, indicating a higher risk associated with this dimension, which aligns with the pathophysiological characteristics of hemodynamic instability caused by abnormal catecholamine secretion in PPGL patients; for the laboratory test dimension, the fields of this dimension reflect the patient's endocrine abnormality level, organ function status, and internal environment stability, thus affecting... The calculation of risk factor values ​​is influenced by various factors. Higher levels of catecholamines and their metabolites, more severe electrolyte disturbances, and more significant organ dysfunction result in higher risk factor values ​​for the corresponding fields and lower risk scores for the dimensions, indicating a higher risk associated with that dimension, corresponding to the pathophysiological characteristics of the PPGL. Regarding the preoperative intervention and medication dimension: fields in this dimension reflect the adequacy of preoperative risk control and the coverage of drug intervention, affecting the calculation of risk factor values. Insufficient preoperative drug intervention, inadequate medication duration, and non-standard dosage adjustments result in higher risk factor values ​​for the corresponding fields and lower risk scores for the dimensions, indicating a higher risk associated with that dimension, corresponding to the clinical diagnostic and treatment requirements for preoperative preparation in the PPGL. Regarding the perioperative planning dimension: fields in this dimension reflect the degree of surgical trauma, the difficulty of anesthesia management, and the expected level of intraoperative risk exposure, affecting the calculation of risk factor values. Greater surgical trauma, longer expected surgical duration, and higher expected blood loss result in higher risk factor values ​​for the corresponding fields and lower risk scores for the dimensions, indicating a higher risk associated with that dimension.

[0046] Determine the dynamic weights of all fields included in a given dimension d. Then, the risk score corresponding to dimension d can be calculated using the following formula:

[0047] in, Let d be the risk score corresponding to dimension d of the i-th patient; The risk factor value for the j-th standard field of the i-th patient is [0,1], which is used to characterize the risk level corresponding to this field. The higher the value, the higher the corresponding risk. The dynamic weight of the j-th field for the i-th patient. This is the set of standard fields corresponding to dimension d.

[0048] It should be noted that, Different calculation rules apply to different field types. For numerical fields, the calculation is based on the deviation of the standardized value of the field from the risk threshold. For positive risk fields (the higher the value, the higher the risk, such as systolic blood pressure and catecholamine levels), the calculation is based on the deviation of the standardized value of the field from the risk threshold. ,in For the j-th standard field of the i-th patient, the unit standardization and min The standardized value obtained after max normalization has a value range of [0,1]. For negative risk fields (the lower the value, the higher the risk, such as hemoglobin level and blood potassium level). For categorization fields, values ​​are predefined based on the risk level corresponding to the classification code; the higher the risk level, the higher the risk level. The closer the value is to 1; for time-series feature fields, the risk level is calculated based on the fluctuation characteristics, such as the higher the coefficient of variation of blood pressure. The closer the value is to 1.

[0049] S203, based on multiple risk scores of the i-th patient, determine the preoperative risk score of the i-th patient.

[0050] Specifically, the preoperative risk score of patient i can be determined based on the first risk score, second risk score, third risk score, fourth risk score, and fifth risk score of patient i.

[0051] In the embodiments provided in this application, before executing step S203, the clinical core score of the d-th dimension, the average confidence score of all fields in the d-th dimension of the i-th patient, and the average population information gain of all fields in the d-th dimension of the i-th patient can be obtained, thereby determining the weight of the d-th dimension of the i-th patient based on these values; then, the weights of the i-th patient in each dimension can be used to perform a weighted summation of the first risk score, second risk score, third risk score, fourth risk score, and fifth risk score of the i-th patient to obtain the preoperative risk score of the i-th patient.

[0052] Weights of the d-th dimension for the i-th patient The specific method for determining it can be calculated using the following formula:

[0053] Where η, θ, and μ are predefined dimension weight hyperparameters, satisfying η+θ+μ=1; The clinical core score of the d-th dimension is assigned a value based on the predefined clinical guidelines and expert consensus on preoperative risk management of PPGL, with a value range of [0,1], which is used to characterize the overall impact of this dimension on preoperative risk of PPGL. Let be the average confidence score of all fields within the d-th dimension for the i-th patient, calculated using the following formula:

[0054] in, The total number of fields in the d-th dimension. This is the set of standard fields corresponding to the d-th dimension; The average population information gain value of all fields within the d-th dimension for the i-th patient is calculated using the following formula:

[0055] After determining the weight of the d-th dimension for the i-th patient, the preoperative risk score for the i-th patient can be determined based on the risk scores of each dimension obtained in step S202 and their corresponding dynamic weights. The specific calculation formula is as follows:

[0056] in, The weight of the d-th dimension for the i-th patient obtained from the above steps; The dimension risk score for the i-th patient dimension d is the first risk score, second risk score, third risk score, fourth risk score, or fifth risk score.

[0057] To address the shortcomings of existing technologies in adapting to individual differences, this application designs a two-layer dynamic weighting calculation mechanism at the field and dimension levels. The weighting calculation process combines multiple factors such as the clinical relevance of the field, population information gain, data confidence, and time validity. It can adjust the weight allocation in real time according to the individual data of each patient, which can adapt to the clinical characteristics of PPGL patients with large fluctuations in indicators and significant individual differences, thereby improving the rationality and adaptability of risk score calculation.

[0058] S204, if the preoperative risk score of the i-th patient is greater than or equal to the risk threshold of the i-th patient, then the prediction result is that the i-th patient has no preoperative anesthesia risk.

[0059] In the embodiments provided in this application, the risk threshold of the i-th patient can be determined in the following way: First, obtain the basic risk threshold, the average confidence level of all fields, and the basic risk factor of the i-th patient; then, determine the risk threshold of the i-th patient based on the basic risk threshold, the average confidence level of all fields, and the basic risk factor of the i-th patient.

[0060] Specifically, the dynamic risk assessment threshold for the i-th patient can be determined using the following formula. Perform the calculation:

[0061] in, The basic threshold can be predefined based on the risk score distribution and risk event occurrence of the PPGL patient historical dataset, with a value range of [0,1]. , These are predefined threshold adjustment coefficients, all of which are greater than 0; The average confidence score for all fields of the target patient is calculated using the following formula:

[0062] in, The total number of all standard fields. The baseline risk factors for the target patients are predefined based on age, ASA classification, and tumor malignancy, with values ​​ranging from [0,1]. The higher the baseline risk, the better. The higher the value, the better.

[0063] Finally, the preoperative risk score of the i-th patient will be... The dynamic risk assessment threshold for the i-th patient In comparison, ≥ At that time, the prediction result is that the i-th patient does not have preoperative anesthesia risk.

[0064] To address the shortcomings of existing technologies that have fixed risk assessment thresholds and are prone to misjudgment, this application designs a dynamic risk assessment threshold calculation mechanism. The threshold can be adjusted according to the patient's data quality and basic risk level, thereby improving the accuracy of risk status identification and reducing the occurrence of false negative and false positive results.

[0065] Based on the above description, this application has the following beneficial effects: This application obtains multi-dimensional perioperative data for the i-th patient, including: demographic and medical history data, physiological sign time-series data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. Based on the demographic and medical history data, a first risk score is determined for the i-th patient; based on the physiological sign time-series data, a second risk score is determined; based on the laboratory test data, a third risk score is determined; based on the preoperative intervention and medication data, a fourth risk score is determined; and based on the perioperative planning data, a fifth risk score is determined. Based on the first, second, third, fourth, and fifth risk scores, a preoperative risk score for the i-th patient is determined. If the preoperative risk score for the i-th patient is greater than or equal to the patient's risk threshold, a prediction result is obtained that the i-th patient has no preoperative anesthesia risk. This invention effectively addresses the problems of inaccurate predictions, difficulty in adapting to individual differences, and lack of attribution guidance caused by existing technologies that rely on physicians' subjective experience and single-dimensional scoring systems. This application collects data from five dimensions: demographics and medical history, physiological signs and their time sequence, laboratory tests, preoperative intervention and medication, and perioperative planning. Corresponding risk scores are calculated for each dimension and then integrated to obtain an individualized preoperative risk score. This score is then combined with patient-specific risk thresholds for risk assessment, achieving a shift from subjective experience-based judgment to objective quantitative evaluation. The multi-dimensional data collection and stratified scoring model comprehensively cover the key risk factors for PPGL patients, improving the comprehensiveness and accuracy of preoperative anesthesia risk assessment for PPGL, while dynamically adapting to individual differences among patients. This reduces the risk of serious perioperative cardiovascular events, ensuring patient safety, and establishes a standardized assessment process, which is beneficial for clinical promotion and anesthesia quality control.

[0066] In the embodiments provided in this application, a prediction result of preoperative anesthesia risk for the i-th patient can also be obtained when the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient.

[0067] It should be noted that when the preoperative risk score of the i-th patient is lower than the risk threshold for the i-th patient, it indicates a higher surgical risk. However, the existing technology can only speculate on the reasons for this higher risk based on expert experience. In this regard, the embodiments provided in this application can also... < When the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient, a prediction result of the preoperative anesthesia risk of the i-th patient is obtained, and the source of the risk can be identified and the contribution can be quantified.

[0068] Specifically, if the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient, the contribution of each dimension of the i-th patient to the overall risk and the contribution of each field in each dimension to the overall risk are calculated; then, a risk assessment report is generated based on the contribution of each dimension of the i-th patient to the overall risk and the contribution of each field in each dimension to the overall risk.

[0069] Wherein, the contribution of each dimension of the i-th patient to the overall risk The following formula can be used for calculation:

[0070] in, Let d be the contribution of the d-th dimension to the overall risk, with a value range of [0,1]. It should be noted that the sum of the contributions of all dimensions is 1. Let d be the absolute risk value of the d-th dimension. The higher the value, the higher the risk brought by that dimension. Let d be the dimension-level dynamic weight of the d-th dimension; For the i-th patient and the k-th dimension, the dimension-level dynamic weights are used. The dimensional risk score for the i-th patient and the k-th dimension.

[0071] To calculate the contribution of each field in each dimension to the overall risk, we can calculate the field-level risk contribution of each field within the top N dimensions (where N is a predefined positive integer) ranked from highest to lowest in terms of dimension-level risk contribution. The specific calculation formula is as follows:

[0072] in, The risk contribution of the j-th field within its respective dimension ranges from [0,1], and the sum of the contributions of all fields within the same dimension is 1. The field-level dynamic weight of the j-th field; The risk factor value for the j-th field; This is the set of standard fields corresponding to this dimension; For the i-th patient, the field-level dynamic weight of the m-th field within the same dimension; For the i-th patient, the risk factor value of the m-th field within the same dimension.

[0073] Finally, based on the dimensional risk contribution... Contribution to Field-Level Risk The calculation results are sorted from highest to lowest contribution and a risk attribution report is generated. The report includes the dimension of the risk source, the corresponding field, the original value of the field, the risk factor value, and the contribution percentage.

[0074] To address the shortcomings of existing technologies in locating the source of risk, this application designs a two-tiered risk attribution quantification mechanism at the dimension and field levels. After determining the existence of risk, it can clearly quantify the contribution of each dimension and field to the overall risk, providing targeted information for clinical intervention and enhancing the clinical practical value of the plan.

[0075] In the embodiments provided in this application, perioperative multidimensional data, preoperative risk scores, and prediction results for the i-th patient can also be stored. This ensures the consistency and reusability of multidimensional data, and the standardized storage based on individual patients solves the problem of multiple preoperative assessments and scattered multi-source heterogeneous data for PPGL patients. It enables unified management and reuse of assessment data from all previous assessments of the same patient, avoids repeated collection and calculation, and improves the operating efficiency of the prediction method.

[0076] It should be noted that the storage of data in this application must comply with the legal obligations for the retention of medical data and personal information, and must strictly meet the legal requirements for the retention and traceability of sensitive medical personal information. The data source, processing procedure, output results, and scope of use must be fully recorded to provide a complete compliance chain for regulatory audits. At the same time, the compliance principle of "minimum necessity and limited authorization" must be implemented, with the stored content strictly limited to information necessary for PPGL preoperative risk prediction, and no collection or storage exceeding the scope allowed. Furthermore, the entire data processing chain can be recorded to verify that all data processing actions are within the scope of the patient's individual written authorization, thus ensuring the legality of data processing from a procedural perspective.

[0077] The above text combined Figure 1 The PPGL preoperative anesthesia risk prediction method provided in the embodiments of this application has been described in detail. The device and equipment provided in the embodiments of this application will be described below with reference to the accompanying drawings.

[0078] like Figure 2 As shown in the figure, this is a schematic diagram of a PPGL preoperative anesthesia risk prediction device 300 provided in an embodiment of this application. The device 300 includes: The multi-dimensional data acquisition module 301 is used to acquire perioperative multi-dimensional data of the i-th patient. The perioperative multi-dimensional data includes: demographic and medical history data, physiological signs time sequence data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. The multidimensional risk score determination module 302 is used to determine the first risk score of the i-th patient based on demographic and medical history data, the second risk score of the i-th patient based on physiological signs time sequence data, the third risk score of the i-th patient based on laboratory test data, the fourth risk score of the i-th patient based on preoperative intervention and medication data, and the fifth risk score of the i-th patient based on perioperative planning data. The preoperative risk score determination module 303 is used to determine the preoperative risk score of the i-th patient based on the first risk score, second risk score, third risk score, fourth risk score and fifth risk score of the i-th patient; The prediction result determination module 304 is used to obtain a prediction result that the i-th patient has no preoperative anesthesia risk when the preoperative risk score of the i-th patient is greater than or equal to the risk threshold of the i-th patient.

[0079] Optionally, the prediction result determination module 304 is also used to obtain a prediction result that the i-th patient has a preoperative anesthesia risk when the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient.

[0080] Optionally, the PPGL preoperative anesthesia risk prediction device 300 also includes: The dimension weight determination module is used to obtain the clinical core score of the d-th dimension, the average confidence score of all fields in the d-th dimension of the i-th patient, and the average population information gain score of all fields in the d-th dimension of the i-th patient, and to determine the weight of the d-th dimension of the i-th patient. The preoperative risk score determination module 303 is also used to use the weights of the i-th patient in each dimension to perform a weighted summation of the first risk score, second risk score, third risk score, fourth risk score and fifth risk score of the i-th patient to obtain the preoperative risk score of the i-th patient.

[0081] Optionally, the prediction result determination module 304 includes a risk threshold determination unit, used to obtain the basic risk threshold, the average confidence level of all fields, and the basic risk factor for the i-th patient; and to determine the risk threshold for the i-th patient based on the basic risk threshold, the average confidence level of all fields, and the basic risk factor for the i-th patient.

[0082] Optionally, the prediction result determination module 304 includes a risk assessment report generation unit, which is used to calculate the contribution of each dimension of the i-th patient to the overall risk and the contribution of each field in each dimension to the overall risk when the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient; and generate a risk assessment report based on the contribution of each dimension of the i-th patient to the overall risk and the contribution of each field in each dimension to the overall risk.

[0083] Optionally, the multi-dimensional data acquisition module 301 is further configured to acquire the initial multi-dimensional data of the i-th patient; and to perform standardization processing on the initial multi-dimensional data of the i-th patient to obtain the perioperative multi-dimensional data of the i-th patient.

[0084] Optionally, the PPGL preoperative anesthesia risk prediction device 300 also includes: The storage module is used to store the perioperative multidimensional data, preoperative risk score, and prediction results of the i-th patient.

[0085] The PPGL preoperative anesthesia risk prediction device according to the embodiments of this application can correspond to the execution of the method described in the embodiments of this application, and the other operations and / or functions of each module / unit of the PPGL preoperative anesthesia risk prediction device are for the purpose of achieving Figure 1 For the sake of brevity, the corresponding processes of each method in the illustrated embodiments will not be described in detail here.

[0086] This application also provides a computing device. For example... Figure 3 As shown in the figure, this is a schematic diagram of a computing device provided in an embodiment of this application. The computing device 700 includes a bus 701, a processor 702, a communication interface 703, and a memory 704. The processor 702, the memory 704, and the communication interface 703 communicate with each other via the bus 701.

[0087] The 701 bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0088] The processor 702 can be any one or more of the following processors: central processing unit (CPU), graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0089] The communication interface 703 is used for external communication.

[0090] Memory 704 may include volatile memory, such as random access memory (RAM). Memory 704 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0091] The memory 704 stores executable code, and the processor 702 executes the executable code to perform the aforementioned PPGL preoperative anesthesia risk prediction method.

[0092] Specifically, in achieving Figure 2 In the case of the illustrated embodiment, and Figure 2 When the modules or units of the PPGL preoperative anesthesia risk prediction device described in the embodiments are implemented by software, the following steps are performed: Figure 2 The software or program code required for the functions of each module / unit can be partially or entirely stored in memory 704. Processor 702 executes the program code corresponding to each unit stored in memory 704 to execute the aforementioned PPGL preoperative anesthesia risk prediction method.

[0093] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to execute the aforementioned PPGL preoperative anesthesia risk prediction method.

[0094] This application also provides a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application are generated.

[0095] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another website, computer, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0096] When the computer program product is executed by a computer, the computer performs any of the methods described in the aforementioned PPGL preoperative anesthesia risk prediction method. The computer program product can be a software installation package; when any of the aforementioned PPGL preoperative anesthesia risk prediction methods needs to be used, the computer program product can be downloaded and executed on the computer.

[0097] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0098] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered within the scope of protection of this application.

Claims

1. A method for predicting the risk of preoperative anesthesia for PPGL, characterized by, The method includes: Obtain perioperative multidimensional data for the i-th patient. The perioperative multidimensional data includes: demographic and medical history data, physiological signs time sequence data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. Based on demographic and medical history data, the first risk score of the i-th patient is determined; based on physiological signs and time series data, the second risk score of the i-th patient is determined; based on laboratory test data, the third risk score of the i-th patient is determined; based on preoperative intervention and medication data, the fourth risk score of the i-th patient is determined; and based on perioperative planning data, the fifth risk score of the i-th patient is determined. Based on the first risk score, second risk score, third risk score, fourth risk score and fifth risk score of the i-th patient, determine the preoperative risk score of the i-th patient; If the preoperative risk score of the i-th patient is greater than or equal to the risk threshold of the i-th patient, then the prediction result is that the i-th patient has no preoperative anesthesia risk.

2. The method of claim 1, wherein, The method further includes: If the preoperative risk score of the i-th patient is less than the risk threshold of the i-th patient, then the prediction result of the i-th patient having preoperative anesthesia risk is obtained.

3. The method of claim 1, wherein, The method further includes: Obtain the clinical core score of the d-th dimension, the average confidence score of all fields in the d-th dimension for the i-th patient, and the average population information gain score of all fields in the d-th dimension for the i-th patient, and determine the weight of the d-th dimension for the i-th patient. The preoperative risk score for the i-th patient is determined based on the first, second, third, fourth, and fifth risk scores, including: Using the weights of the i-th patient in each dimension, the first risk score, second risk score, third risk score, fourth risk score, and fifth risk score of the i-th patient are weighted and summed to obtain the preoperative risk score of the i-th patient.

4. The method of claim 1, wherein, The risk threshold for the i-th patient is determined in the following manner: Obtain the baseline risk threshold, the average confidence level of all fields, and the baseline risk factor for the i-th patient; The risk threshold for patient i is determined based on the baseline risk threshold for patient i, the average confidence level of all fields, and the baseline risk factor.

5. The method of claim 2, wherein, The method further includes: If the preoperative risk score of patient i is less than the risk threshold of patient i, calculate the contribution of each dimension of patient i to the overall risk and the contribution of each field in each dimension to the overall risk. A risk assessment report is generated based on the contribution of each dimension of the i-th patient to the overall risk and the contribution of each field in each dimension to the overall risk.

6. The method of claim 1, wherein, The acquisition of perioperative multidimensional data for the i-th patient includes: Obtain the initial multidimensional data for the i-th patient; The initial multidimensional data of the i-th patient are standardized to obtain the perioperative multidimensional data of the i-th patient.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Store the perioperative multidimensional data, preoperative risk score, and prediction results for the i-th patient.

8. A PPGL preoperative anesthesia risk prediction device, characterized in that, include: The multi-dimensional data acquisition module is used to acquire perioperative multi-dimensional data for the i-th patient. The perioperative multi-dimensional data includes: demographic and medical history data, physiological signs time series data, laboratory test data, preoperative intervention and medication data, and perioperative planning data. The multidimensional risk scoring module is used to determine the first risk score of the i-th patient based on demographic and medical history data, the second risk score of the i-th patient based on physiological signs time sequence data, the third risk score of the i-th patient based on laboratory test data, the fourth risk score of the i-th patient based on preoperative intervention and medication data, and the fifth risk score of the i-th patient based on perioperative planning data. The preoperative risk score determination module is used to determine the preoperative risk score of the i-th patient based on the first risk score, second risk score, third risk score, fourth risk score, and fifth risk score of the i-th patient. The prediction result determination module is used to obtain a prediction result that the i-th patient has no preoperative anesthesia risk when the preoperative risk score of the i-th patient is greater than or equal to the risk threshold of the i-th patient.

9. A computing device, characterized in that, Including memory and processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program for performing the method as described in any one of claims 1 to 7.

Images