Anesthesia drug efficacy dynamic evaluation system and method based on artificial intelligence

Abstract

The application discloses an anesthetic efficacy dynamic evaluation system and method based on artificial intelligence, relates to the technical field of anesthetic efficacy evaluation, and comprises a simulation prediction unit, a comparison correction unit, an abnormal backtracking unit and an output warning unit; the simulation prediction unit is used for collecting real-time physiological parameters of a patient through a medical monitoring device, constructing a virtual anesthetic mirror according to the real-time physiological parameters, generating a virtual physiological model through a drug metabolism dynamics algorithm, outputting anesthetic efficacy evaluation data and sending the anesthetic efficacy evaluation data to the comparison correction unit, and can reduce the occurrence of deep or shallow anesthesia caused by inaccurate evaluation, reduce the operation risk, guarantee the life safety of the patient, enable an anesthesiologist to know possible anesthesia state changes of the patient in advance, timely adjust an anesthesia scheme, such as adjusting the infusion rate, the dose and the like of an anesthetic, ensure the stability of the anesthesia state in the operation process, improve the smooth operation degree of the operation, and provide more comprehensive and detailed decision basis for the anesthesiologist.

Description

Technical Field

[0001] This invention relates to the field of anesthetic efficacy evaluation technology, specifically to an artificial intelligence-based dynamic evaluation system and method for anesthetic efficacy. Background Technology

[0002] Anesthesia is the temporary loss of sensation in a patient, either whole or locally, through the use of drugs or other methods to achieve painlessness and create favorable conditions for surgical treatment or other medical examinations. In the field of medical anesthesia, accurate assessment of the efficacy of anesthetic drugs is crucial for the success of surgery and the safety of patients.

[0003] Traditional methods for assessing the efficacy of anesthetic drugs mainly rely on the clinical experience of anesthesiologists and simple monitoring of some basic physiological indicators. Doctors usually judge the appropriate depth of anesthesia by observing the patient's vital signs, such as heart rate, blood pressure, and respiratory rate, combined with their own experience. However, the human physiological system is extremely complex, and different patients respond significantly differently to anesthetic drugs. Even the same dose and type of anesthetic drug will not produce the same effect in different patients. Relying solely on limited basic physiological indicators and physician experience is difficult to comprehensively and accurately assess the efficacy of anesthetic drugs, which can easily lead to situations where the anesthesia is too deep or too shallow. If the anesthesia is too deep, it can cause a series of serious complications, such as respiratory depression and circulatory depression; while if the anesthesia is too shallow, it can cause the patient to experience movement and loss of consciousness during surgery.

[0004] Therefore, in view of this, the present invention proposes an artificial intelligence-based dynamic evaluation system and method for anesthetic efficacy to make up for and improve the shortcomings of the existing technology. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides an artificial intelligence-based dynamic evaluation system and method for anesthetic efficacy, thereby resolving the corresponding technical issues raised in the background section.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is: a dynamic evaluation system for anesthetic efficacy based on artificial intelligence, including a simulation prediction unit, a comparison and correction unit, an anomaly backtracking unit, and an output early warning unit;

[0007] The simulation prediction unit is used to collect the patient's real-time physiological parameters through medical monitoring equipment, construct a virtual anesthesia image based on the real-time physiological parameters, generate a virtual physiological model through pharmacokinetic algorithm, output anesthetic efficacy evaluation data and send it to the comparison and correction unit.

[0008] The comparison and correction unit is used to acquire anesthetic efficacy assessment data and the patient's actual physiological parameters during the operation. It compares the actual physiological parameters with the anesthetic efficacy assessment data, calculates the deviation value, adjusts the virtual physiological model based on the reinforcement learning algorithm, and generates updated anesthetic efficacy assessment data based on the adjusted virtual physiological model and sends it to the anomaly backtracking unit.

[0009] The anomaly backtracking unit is used to acquire updated anesthetic efficacy assessment data, predict the trend of changes in the patient's anesthetic depth within a preset time period based on the adjusted virtual physiological model, analyze and determine the patient's physiological abnormalities in combination with the preset conventional trend of anesthetic efficacy, simultaneously acquire the abnormal time corresponding to the physiological abnormality and the actual physiological parameters corresponding to the abnormal time, perform backtracking analysis on the physiological abnormality, determine the impact of anesthetic drugs on the physiological abnormality, and send the determination results to the output warning unit.

[0010] The output warning unit is used to obtain a definite result, compare it with a preset anesthetic efficacy safety threshold, classify different warning levels according to the comparison result, generate warning information and send it to the physician terminal.

[0011] Preferably, the real-time physiological parameters include heart rate (AA), blood oxygen saturation (AB), systolic blood pressure (AC), diastolic blood pressure (AD), and respiratory rate (AE), and the anesthetic efficacy assessment data include depth of anesthesia (BIS), volume of distribution (Vd), blood drug concentration (C), and anesthetic efficacy assessment index (effect).

[0012] As a preferred method, the specific process for outputting anesthetic efficacy evaluation data is as follows:

[0013] S101. Based on the patient's real-time physiological parameters, construct the patient's physiological feature vector V, the formula of which is:

[0014] ;

[0015] in, As weighted by heart rate, Weighted by blood oxygen saturation. As the weight of systolic blood pressure, As the weight of diastolic blood pressure, Weighted by respiratory rate;

[0016] S102. Based on the long short-term memory network model, establish an anesthesia state mapping model, and use the anesthesia state mapping model as the basic model of the virtual anesthesia mirror. Obtain the physiological parameters and anesthesia state information of historical anesthetized patients from the medical database and integrate them into a historical dataset. Divide the historical dataset into a training set and a validation set in a 7:3 ratio. Use the training set to train the basic model and adjust the internal parameters of the neural network. Use the validation set to validate the trained basic model.

[0017] S103. Input the patient's physiological feature vector V into the basic model to obtain the anesthesia depth BIS, the formula of which is:

[0018] ;

[0019] Where f is the nonlinear mapping function of the basic model, and θ is the bias term;

[0020] S104. Obtain the patient's weight BW, generate a virtual physiological model using a pharmacokinetic algorithm, and estimate the volume of distribution Vd of the anesthetic drug in the patient's body based on the patient's weight using an empirical formula:

[0021] ;

[0022] Where a and b are empirical constants for anesthetic drugs;

[0023] S105. Based on the one-compartment model of pharmacokinetics, calculate the blood concentration of the anesthetic drug in the patient. Assuming the anesthetic drug is infused at a constant rate and the elimination rate constant is k, the formula for the change in blood concentration over time is:

[0024] ;

[0025] Where t is time, Let be the infusion rate constant of the anesthetic drug. This refers to the constant infusion rate parameter for anesthetic drugs;

[0026] S106. Based on the pharmacodynamic characteristics of anesthetic drugs, establish a linear conversion relationship between blood drug concentration and anesthetic efficacy evaluation indicators. The formula is as follows:

[0027] ;

[0028] Wherein, effect is the anesthetic efficacy evaluation index, C is the blood drug concentration, m is the slope, and c is the intercept.

[0029] As a preferred method, the specific process for generating updated anesthetic efficacy evaluation data is as follows:

[0030] S201. Obtain anesthetic efficacy assessment data and simultaneously obtain the patient's actual physiological parameters during the operation. Use linear interpolation to align the actual physiological parameters with the anesthetic efficacy assessment data over time as follows:

[0031] The set of actual physiological parameters at each time point t With anesthetic efficacy evaluation data set Pairing data point by point to form data pairs ;

[0032] S202. For each time point t, calculate the deviation value parameter by parameter. as follows:

[0033] The deviation of real-time physiological parameters is normalized using the clinical safety range, and the deviation formula is as follows:

[0034] ;

[0035] in, , This is the threshold for clinical safety. These are the parameter weighting coefficients;

[0036] A deviation function was designed based on pharmacokinetic characteristics, and an asymmetric error penalty function was used to calculate the deviation value of the Biological Intensity Sequence (BIS) of anesthesia depth. The deviation formula is as follows:

[0037] ;

[0038] in, , For hyperparameters;

[0039] The deviation value of blood drug concentration C is calculated using relative deviation, and the deviation formula is as follows:

[0040] ;

[0041] in, This is the concentration sensitivity coefficient;

[0042] The deviation between the volume of distribution Vd and the anesthetic efficacy evaluation index effect is calculated using the absolute deviation method. The deviation formula is as follows:

[0043] ;

[0044] ;

[0045] in, The weights of the distribution volume, The weights of the anesthetic efficacy evaluation indicators;

[0046] The deviation values ​​of all parameters constitute the deviation vector. The overall deviation value is generated by weighted summation. ,in, Weights for each parameter;

[0047] S203. The generated comprehensive deviation value is used as the input feature vector of the reinforcement learning algorithm. The Q-learning algorithm is adopted, defining the state as the parameter set of the current virtual physiological model and the action as the adjustment of the virtual physiological model parameters. When the deviation value between the new anesthetic efficacy evaluation data and the actual physiological parameters decreases after the virtual physiological model is adjusted, a positive reward is given, and a negative reward is given when the deviation value increases. Through multiple iterations of training, the virtual physiological model parameters are optimized. Based on the adjusted virtual physiological model, the anesthesia depth, distribution volume, blood drug concentration and anesthetic efficacy evaluation index are recalculated to generate updated anesthetic efficacy evaluation data.

[0048] As a preferred method, the specific process for predicting the trend of changes in the patient's depth of anesthesia within a predetermined time period is as follows:

[0049] S301. Obtain the updated anesthetic efficacy assessment data. Using the current time point as a baseline, set a prediction time window. Based on a long short-term memory network model, use the updated anesthetic efficacy assessment data as input to generate a sequence prediction value of the change in anesthesia depth over time. The formula is as follows:

[0050] ;

[0051] in, For the predicted value of the depth of anesthesia at the i-th time point in the future, the input features include the current depth of anesthesia, blood drug concentration, volume of distribution, and anesthetic efficacy evaluation indicators.

[0052] S302. Based on the generated sequence prediction value of the change of anesthesia depth over time, with time as the horizontal axis and anesthesia depth BIS as the vertical axis, output the trend curve of the change of the patient's anesthesia depth within a future preset time period.

[0053] As a preferred method, the specific process for determining the effects of anesthetic drugs on physiological abnormalities is as follows:

[0054] S401. Based on the pre-set trend of anesthetic efficacy, analyze and identify the patient's physiological abnormalities. Simultaneously, when identifying the physiological abnormality, obtain the corresponding time of the abnormality, the actual physiological parameters at that time, and the corresponding anesthetic efficacy assessment data. Calculate the correlation index r between the actual physiological parameters and the anesthetic efficacy assessment data using the Pearson correlation coefficient. The formula is as follows:

[0055] ;

[0056] S402. Based on the preset correlation threshold, the calculated correlation index r is compared with the preset correlation threshold. Comparisons were performed, and retrospective analysis was conducted on physiological abnormalities. Therefore, anesthetic drugs are identified as the main influencing factor of physiological abnormalities.

[0057] As a preferred method, the specific process for generating early warning information is as follows:

[0058] S501. Obtain the confirmed results and calculate the Abnormality Severity Index (ASI) caused by anesthetic drugs to physiological abnormalities. The formula is as follows:

[0059] ;

[0060] Wherein, Im represents the degree of influence of the anesthetic drug, which is obtained through conversion using the correlation index r. The weights of the parameter deviation values. The weighting of the degree of influence of anesthetic drugs;

[0061] S502. Based on the preset safety threshold for anesthetic efficacy, the warning levels are divided according to the Abnormal Severity Index (ASI) as follows:

[0062] when When this occurs, it is classified as a Level 1 warning, requiring emergency intervention to adjust the anesthetic dosage;

[0063] when At that time, it was classified as a Level II early warning system, and the anesthetic dosage was closely monitored;

[0064] when At that time, it was divided into three levels of early warning, and the anesthetic dosage was routinely monitored.

[0065] An artificial intelligence-based method for dynamic evaluation of anesthetic efficacy includes the following steps:

[0066] Step 1: Collect the patient's real-time physiological parameters through medical monitoring equipment, construct a virtual anesthesia image based on the real-time physiological parameters, generate a virtual physiological model through pharmacokinetic algorithms, and output anesthetic efficacy assessment data;

[0067] Step 2: Obtain anesthetic efficacy assessment data and the patient's actual physiological parameters during the operation. Compare the actual physiological parameters with the anesthetic efficacy assessment data, calculate the deviation value, adjust the virtual physiological model based on the reinforcement learning algorithm, and generate updated anesthetic efficacy assessment data based on the adjusted virtual physiological model.

[0068] Step 3: Obtain updated anesthetic efficacy assessment data. Based on the adjusted virtual physiological model, predict the trend of changes in the patient's anesthesia depth within a preset time period. Combined with the preset conventional trend of anesthetic efficacy changes, analyze and determine the patient's physiological abnormalities. Simultaneously obtain the abnormal time corresponding to the physiological abnormality and the actual physiological parameters corresponding to the abnormal time. Perform retrospective analysis on the physiological abnormalities to determine the impact of anesthetic drugs on the physiological abnormalities.

[0069] Step 4: Obtain the confirmed results, compare them with the preset safety thresholds for anesthetic efficacy, classify different warning levels based on the comparison results, generate warning information and send it to the physician's end.

[0070] Compared with existing technologies, the beneficial effects of this invention are as follows: By collecting real-time physiological parameters of the patient to construct a virtual anesthesia mirror, generating a virtual physiological model, and outputting anesthetic efficacy assessment data, the invention simultaneously acquires the patient's actual physiological parameters during surgery, calculates deviation values, adjusts the virtual physiological model, and generates updated anesthetic efficacy assessment data based on the adjusted virtual physiological model. It predicts the trend of changes in the patient's anesthesia depth within a preset time period, analyzes and identifies physiological abnormalities in the patient by combining these with preset conventional trends in anesthetic efficacy, and performs retrospective analysis on these abnormalities to determine the impact of anesthetic drugs on these abnormalities. It also compares these abnormalities with preset anesthetic efficacy safety thresholds, classifies different warning levels based on the comparison results, generates warning information, and sends it to the physician. This approach fully considers individual differences and real-time physiological states of patients, thereby providing a more accurate assessment. By assessing the efficacy of anesthetic drugs, anesthesiologists can reduce the occurrence of overly deep or shallow anesthesia due to inaccurate assessments, thereby lowering surgical risks and ensuring patient safety. Predicting the future trend of anesthetic drug efficacy allows anesthesiologists to anticipate potential changes in the patient's anesthetic state and adjust the anesthesia plan accordingly, such as adjusting the infusion rate and dosage of anesthetic drugs. This ensures stable anesthetic status during surgery, improving the success rate of the procedure. Comprehensive analysis of predictive data and actual physiological parameters enables the timely detection of physiological abnormalities during anesthesia and accurate assessment of the impact of anesthetic drugs on these abnormalities. This provides anesthesiologists with more comprehensive and detailed decision-making support, facilitating rapid and accurate implementation of appropriate treatment measures, improving the patient's physiological state, reducing complications, and enhancing the patient's surgical experience and postoperative recovery quality. Attached Figure Description

[0071] Figure 1 This is a schematic diagram of the overall structure of a preferred embodiment of the present invention;

[0072] Figure 2 This is a flowchart of the method shown in this invention. Detailed Implementation

[0073] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] Embodiments of the present invention:

[0075] Please refer to Figures 1 to 2 As shown, an artificial intelligence-based dynamic evaluation system for anesthetic efficacy includes a simulation prediction unit, a comparison and correction unit, an anomaly backtracking unit, and an output early warning unit.

[0076] The simulation prediction unit is used to collect the patient's real-time physiological parameters through medical monitoring equipment, construct a virtual anesthesia image based on the real-time physiological parameters, generate a virtual physiological model through pharmacokinetic algorithms, output anesthetic efficacy assessment data, and send it to the comparison and correction unit.

[0077] The comparison and correction unit is used to acquire anesthetic efficacy assessment data and the patient's actual physiological parameters during the operation. It compares the actual physiological parameters with the anesthetic efficacy assessment data, calculates the deviation value, adjusts the virtual physiological model based on the reinforcement learning algorithm, and generates updated anesthetic efficacy assessment data based on the adjusted virtual physiological model and sends it to the anomaly backtracking unit.

[0078] The anomaly backtracking unit is used to acquire updated anesthetic efficacy assessment data. Based on the adjusted virtual physiological model, it predicts the trend of changes in the patient's anesthetic depth within a preset time period. Combined with the preset routine trend of anesthetic efficacy, it analyzes and determines the patient's physiological abnormalities. Simultaneously, it acquires the abnormal time corresponding to the physiological abnormality and the actual physiological parameters corresponding to the abnormal time. It performs backtracking analysis on the physiological abnormalities to determine the impact of anesthetic drugs on the physiological abnormalities and sends the determination results to the output warning unit.

[0079] The output warning unit is used to obtain a definite result, compare it with the preset anesthetic efficacy safety threshold, classify different warning levels according to the comparison result, generate warning information and send it to the physician.

[0080] Real-time physiological parameters include heart rate (AA), blood oxygen saturation (AB), systolic blood pressure (AC), diastolic blood pressure (AD), and respiratory rate (AE). Anesthetic efficacy assessment data include depth of anesthesia (BIS), volume of distribution (Vd), blood drug concentration (C), and anesthetic efficacy assessment index (effect).

[0081] The specific process for outputting anesthetic efficacy evaluation data is as follows:

[0082] S101. Based on the patient's real-time physiological parameters, construct the patient's physiological feature vector V, the formula of which is:

[0083] ;

[0084] in, As weighted by heart rate, Weighted by blood oxygen saturation. As the weight of systolic blood pressure, As the weight of diastolic blood pressure, Weighted by respiratory rate;

[0085] S102. Based on the long short-term memory network model, establish an anesthesia state mapping model, and use the anesthesia state mapping model as the basic model of the virtual anesthesia mirror. Obtain the physiological parameters and anesthesia state information of historical anesthetized patients from the medical database and integrate them into a historical dataset. Divide the historical dataset into a training set and a validation set in a 7:3 ratio. Use the training set to train the basic model and adjust the internal parameters of the neural network. Use the validation set to validate the trained basic model.

[0086] S103. Input the patient's physiological feature vector V into the basic model to obtain the anesthesia depth BIS, the formula of which is:

[0087] ;

[0088] Where f is the nonlinear mapping function of the basic model, and θ is the bias term;

[0089] S104. Obtain the patient's weight BW, generate a virtual physiological model using a pharmacokinetic algorithm, and estimate the volume of distribution Vd of the anesthetic drug in the patient's body based on the patient's weight using an empirical formula:

[0090] ;

[0091] Where a and b are empirical constants for anesthetic drugs;

[0092] S105. Based on the one-compartment model of pharmacokinetics, calculate the blood concentration of the anesthetic drug in the patient. Assuming the anesthetic drug is infused at a constant rate and the elimination rate constant is k, the formula for the change in blood concentration over time is:

[0093] ;

[0094] Where t is time, Let be the infusion rate constant of the anesthetic drug. This refers to the constant infusion rate parameter for anesthetic drugs;

[0095] S106. Based on the pharmacodynamic characteristics of anesthetic drugs, establish a linear conversion relationship between blood drug concentration and anesthetic efficacy evaluation indicators. The formula is as follows:

[0096] ;

[0097] Wherein, effect is the anesthetic efficacy evaluation index, C is the blood drug concentration, m is the slope, and c is the intercept.

[0098] The specific process for generating updated anesthetic efficacy evaluation data is as follows:

[0099] S201. Obtain anesthetic efficacy assessment data and simultaneously obtain the patient's actual physiological parameters during the operation. Use linear interpolation to align the actual physiological parameters with the anesthetic efficacy assessment data over time as follows:

[0100] The set of actual physiological parameters at each time point t With anesthetic efficacy evaluation data set Pairing data point by point to form data pairs ;

[0101] S202. For each time point t, calculate the deviation value parameter by parameter. as follows:

[0102] The deviation of real-time physiological parameters is normalized using the clinical safety range, and the deviation formula is as follows:

[0103] ;

[0104] in, , This is the threshold for clinical safety. These are the parameter weighting coefficients;

[0105] A deviation function was designed based on pharmacokinetic characteristics, and an asymmetric error penalty function was used to calculate the deviation value of the Biological Intensity Sequence (BIS) of anesthesia depth. The deviation formula is as follows:

[0106] ;

[0107] in, , For hyperparameters;

[0108] The deviation value of blood drug concentration C is calculated using relative deviation, and the deviation formula is as follows:

[0109] ;

[0110] in, This is the concentration sensitivity coefficient;

[0111] The deviation between the volume of distribution Vd and the anesthetic efficacy evaluation index effect is calculated using the absolute deviation method. The deviation formula is as follows:

[0112] ;

[0113] ;

[0114] in, The weights of the distribution volume, The weights of the anesthetic efficacy evaluation indicators;

[0115] The deviation values ​​of all parameters constitute the deviation vector. The overall deviation value is generated by weighted summation. ,in, Weights for each parameter;

[0116] S203. The generated comprehensive deviation value is used as the input feature vector of the reinforcement learning algorithm. The Q-learning algorithm is adopted, defining the state as the parameter set of the current virtual physiological model and the action as the adjustment of the virtual physiological model parameters. When the deviation value between the new anesthetic efficacy evaluation data and the actual physiological parameters decreases after the virtual physiological model is adjusted, a positive reward is given, and a negative reward is given when the deviation value increases. Through multiple iterations of training, the virtual physiological model parameters are optimized. Based on the adjusted virtual physiological model, the anesthesia depth, distribution volume, blood drug concentration and anesthetic efficacy evaluation index are recalculated to generate updated anesthetic efficacy evaluation data.

[0117] The specific process for predicting the trend of changes in a patient's depth of anesthesia within a predetermined time period is as follows:

[0118] S301. Obtain the updated anesthetic efficacy assessment data. Using the current time point as a baseline, set a prediction time window. Based on a long short-term memory network model, use the updated anesthetic efficacy assessment data as input to generate a sequence prediction value of the change in anesthesia depth over time. The formula is as follows:

[0119] ;

[0120] in, For the predicted value of the depth of anesthesia at the i-th time point in the future, the input features include the current depth of anesthesia, blood drug concentration, volume of distribution, and anesthetic efficacy evaluation indicators.

[0121] S302. Based on the generated sequence prediction value of the change of anesthesia depth over time, with time as the horizontal axis and anesthesia depth BIS as the vertical axis, output the trend curve of the change of the patient's anesthesia depth within a future preset time period.

[0122] The specific process for determining the effects of anesthetic drugs on physiological abnormalities is as follows:

[0123] S401. Based on the pre-set trend of anesthetic efficacy, analyze and identify the patient's physiological abnormalities. Simultaneously, when identifying the physiological abnormality, obtain the corresponding time of the abnormality, the actual physiological parameters at that time, and the corresponding anesthetic efficacy assessment data. Calculate the correlation index r between the actual physiological parameters and the anesthetic efficacy assessment data using the Pearson correlation coefficient. The formula is as follows:

[0124] ;

[0125] S402. Based on the preset correlation threshold, the calculated correlation index r is compared with the preset correlation threshold. Comparisons were performed, and retrospective analysis was conducted on physiological abnormalities. Therefore, anesthetic drugs are identified as the main influencing factor of physiological abnormalities.

[0126] The specific process for generating early warning information is as follows:

[0127] S501. Obtain the confirmed results and calculate the Abnormality Severity Index (ASI) caused by anesthetic drugs to physiological abnormalities. The formula is as follows:

[0128] ;

[0129] Wherein, Im represents the degree of influence of the anesthetic drug, which is obtained through conversion using the correlation index r. The weights of the parameter deviation values. The weighting of the degree of influence of anesthetic drugs;

[0130] S502. Based on the preset safety threshold for anesthetic efficacy, the warning levels are divided according to the Abnormal Severity Index (ASI) as follows:

[0131] when When this occurs, it is classified as a Level 1 warning, requiring emergency intervention to adjust the anesthetic dosage;

[0132] when At that time, it was classified as a Level II early warning system, and the anesthetic dosage was closely monitored;

[0133] when At that time, it was divided into three levels of early warning, and the anesthetic dosage was routinely monitored.

[0134] An artificial intelligence-based method for dynamic evaluation of anesthetic efficacy includes the following steps:

[0135] Step 1: Collect the patient's real-time physiological parameters through medical monitoring equipment, construct a virtual anesthesia image based on the real-time physiological parameters, generate a virtual physiological model through pharmacokinetic algorithms, and output anesthetic efficacy assessment data;

[0136] Step 2: Obtain anesthetic efficacy assessment data and the patient's actual physiological parameters during the operation. Compare the actual physiological parameters with the anesthetic efficacy assessment data, calculate the deviation value, adjust the virtual physiological model based on the reinforcement learning algorithm, and generate updated anesthetic efficacy assessment data based on the adjusted virtual physiological model.

[0137] Step 3: Obtain updated anesthetic efficacy assessment data. Based on the adjusted virtual physiological model, predict the trend of changes in the patient's anesthesia depth within a preset time period. Combined with the preset conventional trend of anesthetic efficacy changes, analyze and determine the patient's physiological abnormalities. Simultaneously obtain the abnormal time corresponding to the physiological abnormality and the actual physiological parameters corresponding to the abnormal time. Perform retrospective analysis on the physiological abnormalities to determine the impact of anesthetic drugs on the physiological abnormalities.

[0138] Step 4: Obtain the confirmed results, compare them with the preset safety thresholds for anesthetic efficacy, classify different warning levels based on the comparison results, generate warning information and send it to the physician's end.

[0139] By collecting real-time physiological parameters of patients to construct a virtual anesthesia mirror, a virtual physiological model is generated, and anesthetic efficacy assessment data is output. Simultaneously, the patient's actual intraoperative physiological parameters are acquired, deviation values ​​are calculated, and the virtual physiological model is adjusted. Based on the adjusted virtual physiological model, updated anesthetic efficacy assessment data is generated, predicting the trend of changes in the patient's depth of anesthesia within a preset time period. Combined with preset routine trends in anesthetic efficacy, physiological abnormalities in patients are analyzed and identified. Retrospective analysis of these abnormalities is performed to determine the impact of anesthetic drugs on physiological abnormalities. This is compared with preset anesthetic efficacy safety thresholds, and different warning levels are assigned based on the comparison results. Warning information is generated and sent to the physician's end. This approach fully considers individual patient differences and real-time physiological states, thereby more accurately assessing anesthetic efficacy and reducing errors caused by subjective assessment methods. Inaccurate estimations can lead to anesthesia that is too deep or too shallow, reducing surgical risks and ensuring patient safety. By predicting the changing trends of anesthetic drug efficacy over a future period, anesthesiologists can anticipate potential changes in the patient's anesthetic state and adjust the anesthesia plan in a timely manner, such as adjusting the infusion rate and dosage of anesthetic drugs. This ensures a stable anesthetic state during surgery, improving the success rate of the operation. By comprehensively analyzing predicted data and actual physiological parameters, anesthesiologists can promptly identify physiological abnormalities during anesthesia and accurately assess the extent of the anesthetic drug's impact on these abnormalities. This provides anesthesiologists with a more comprehensive and detailed basis for decision-making, facilitating the rapid and accurate implementation of appropriate treatment measures, improving the patient's physiological state, reducing the occurrence of complications, and enhancing the patient's surgical experience and postoperative recovery quality.

[0140] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0141] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0142] In the two embodiments provided in this application, it should be understood that the disclosed apparatus and system can be implemented in other ways; for example, the apparatus embodiments described above are merely illustrative, and the division of modules is merely a logical functional division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed; furthermore, the coupling or direct coupling or communication connection between the shown or discussed mutuals can be through some interfaces, and the indirect coupling or communication connection between the apparatus or modules can be electrical, mechanical or other forms.

[0143] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A dynamic evaluation system for the efficacy of anesthetic drugs based on artificial intelligence, characterized in that, It includes a simulation prediction unit, a comparison and correction unit, an anomaly backtracking unit, and an output early warning unit; The simulation prediction unit is used to collect the patient's real-time physiological parameters through medical monitoring equipment, construct a virtual anesthesia image based on the real-time physiological parameters, generate a virtual physiological model through pharmacokinetic algorithm, output anesthetic efficacy evaluation data and send it to the comparison and correction unit. The comparison and correction unit is used to acquire anesthetic efficacy assessment data and the patient's actual physiological parameters during the operation. It compares the actual physiological parameters with the anesthetic efficacy assessment data, calculates the deviation value, adjusts the virtual physiological model based on the reinforcement learning algorithm, and generates updated anesthetic efficacy assessment data based on the adjusted virtual physiological model and sends it to the anomaly backtracking unit. The specific process for generating updated anesthetic efficacy evaluation data is as follows: S201. Obtain anesthetic efficacy assessment data and simultaneously obtain the patient's actual physiological parameters during the operation. Use linear interpolation to align the actual physiological parameters with the anesthetic efficacy assessment data over time as follows: The set of actual physiological parameters at each time point t With anesthetic efficacy evaluation data set Pairing data point by point to form data pairs ; S202. For each time point t, calculate the deviation value parameter by parameter. as follows: The deviation of real-time physiological parameters is normalized using the clinical safety range, and the deviation formula is as follows: ； in, , This is the threshold for clinical safety. These are the parameter weighting coefficients; A deviation function was designed based on pharmacokinetic characteristics, and an asymmetric error penalty function was used to calculate the deviation value of the Biological Intensity Sequence (BIS) of anesthesia depth. The deviation formula is as follows: ； in, , For hyperparameters; The deviation value of blood drug concentration C is calculated using relative deviation, and the deviation formula is as follows: ； in, This is the concentration sensitivity coefficient; The deviation between the volume of distribution Vd and the anesthetic efficacy evaluation index effect is calculated using the absolute deviation method. The deviation formula is as follows: ； ； in, The weights of the distribution volume, The weights of the anesthetic efficacy evaluation indicators; The deviation values ​​of all parameters constitute the deviation vector. The overall deviation value is generated by weighted summation. ,in, Weights for each parameter; S203. The generated comprehensive deviation value is used as the input feature vector of the reinforcement learning algorithm. The Q-learning algorithm is adopted, and the state is defined as the parameter set of the current virtual physiological model. The action is to adjust the parameters of the virtual physiological model. When the deviation value between the new anesthetic efficacy evaluation data and the actual physiological parameters decreases after the virtual physiological model is adjusted, a positive reward is given, and a negative reward is given when the deviation value increases. Through multiple iterations of training, the virtual physiological model parameters are optimized. Based on the adjusted virtual physiological model, the anesthesia depth, distribution volume, blood drug concentration and anesthetic efficacy evaluation index are recalculated to generate updated anesthetic efficacy evaluation data. The anomaly backtracking unit is used to acquire updated anesthetic efficacy assessment data, predict the trend of changes in the patient's anesthetic depth within a preset time period based on the adjusted virtual physiological model, analyze and determine the patient's physiological abnormalities in combination with the preset conventional trend of anesthetic efficacy, simultaneously acquire the abnormal time corresponding to the physiological abnormality and the actual physiological parameters corresponding to the abnormal time, perform backtracking analysis on the physiological abnormality, determine the impact of anesthetic drugs on the physiological abnormality, and send the determination results to the output warning unit. The specific process for determining the effects of anesthetic drugs on physiological abnormalities is as follows: S401. Based on the pre-set trend of anesthetic efficacy, analyze and identify the patient's physiological abnormalities. Simultaneously, when identifying the physiological abnormality, obtain the corresponding time of the abnormality, the actual physiological parameters at that time, and the corresponding anesthetic efficacy assessment data. Calculate the correlation index r between the actual physiological parameters and the anesthetic efficacy assessment data using the Pearson correlation coefficient. The formula is as follows: ； S402. Based on the preset correlation threshold, the calculated correlation index r is compared with the preset correlation threshold. Comparisons were performed, and retrospective analysis was conducted on physiological abnormalities. Therefore, anesthetic drugs are identified as the main influencing factor of physiological abnormalities. The output warning unit is used to obtain a definite result, compare it with a preset anesthetic efficacy safety threshold, divide different warning levels according to the comparison result, generate warning information and send it to the physician terminal; The specific process for generating early warning information is as follows: S501. Obtain the confirmed results and calculate the Abnormality Severity Index (ASI) caused by anesthetic drugs to physiological abnormalities. The formula is as follows: ； Wherein, Im represents the degree of influence of the anesthetic drug, which is obtained through conversion using the correlation index r. The weights of the parameter deviation values. The weighting of the degree of influence of anesthetic drugs; S502. Based on the preset safety threshold for anesthetic efficacy, the warning levels are divided according to the Abnormal Severity Index (ASI) as follows: when When this occurs, it is classified as a Level 1 warning, requiring emergency intervention to adjust the anesthetic dosage; when At that time, it was classified as a Level II early warning system, and the anesthetic dosage was closely monitored; when At that time, it was divided into three levels of early warning, and the anesthetic dosage was routinely monitored.

2. The anesthetic efficacy dynamic evaluation system based on artificial intelligence according to claim 1, characterized in that, The real-time physiological parameters include heart rate (AA), blood oxygen saturation (AB), systolic blood pressure (AC), diastolic blood pressure (AD), and respiratory rate (AE). The anesthetic efficacy assessment data include depth of anesthesia (BIS), volume of distribution (Vd), blood drug concentration (C), and anesthetic efficacy assessment index (effect).

3. The dynamic evaluation system for anesthetic efficacy based on artificial intelligence according to claim 2, characterized in that, The specific process for outputting anesthetic efficacy evaluation data is as follows: S101. Based on the patient's real-time physiological parameters, construct the patient's physiological feature vector V, the formula of which is: ； in, As weighted by heart rate, Weighted by blood oxygen saturation. As the weight of systolic blood pressure, As the weight of diastolic blood pressure, Weighted by respiratory rate; S102. Based on the long short-term memory network model, establish an anesthesia state mapping model, and use the anesthesia state mapping model as the basic model of the virtual anesthesia mirror. Obtain the physiological parameters and anesthesia state information of historical anesthetized patients from the medical database and integrate them into a historical dataset. Divide the historical dataset into a training set and a validation set in a 7:3 ratio. Use the training set to train the basic model and adjust the internal parameters of the neural network. Use the validation set to validate the trained basic model. S103. Input the patient's physiological feature vector V into the basic model to obtain the anesthesia depth BIS, the formula of which is: ； Where f is the nonlinear mapping function of the basic model, and θ is the bias term; S104. Obtain the patient's weight BW, generate a virtual physiological model using a pharmacokinetic algorithm, and estimate the volume of distribution Vd of the anesthetic drug in the patient's body based on the patient's weight using an empirical formula: ； Where a and b are empirical constants for anesthetic drugs; S105. Based on the one-compartment model of pharmacokinetics, calculate the blood concentration of the anesthetic drug in the patient. Assuming the anesthetic drug is infused at a constant rate and the elimination rate constant is k, the formula for the change in blood concentration over time is: ； Where t is time, Let be the infusion rate constant of the anesthetic drug. This refers to the constant infusion rate parameter for anesthetic drugs; S106. Based on the pharmacodynamic characteristics of anesthetic drugs, establish a linear conversion relationship between blood drug concentration and anesthetic efficacy evaluation indicators. The formula is as follows: ； Wherein, effect is the anesthetic efficacy evaluation index, C is the blood drug concentration, m is the slope, and c is the intercept.

4. The dynamic evaluation system for anesthetic efficacy based on artificial intelligence according to claim 3, characterized in that, The specific process for predicting the trend of changes in a patient's depth of anesthesia within a predetermined time period is as follows: S301. Obtain the updated anesthetic efficacy assessment data. Using the current time point as a baseline, set a prediction time window. Based on a long short-term memory network model, use the updated anesthetic efficacy assessment data as input to generate a sequence prediction value of the change in anesthesia depth over time. The formula is as follows: ； in, For the predicted value of the depth of anesthesia at the i-th time point in the future, the input features include the current depth of anesthesia, blood drug concentration, volume of distribution, and anesthetic efficacy evaluation indicators. S302. Based on the generated sequence prediction value of the change of anesthesia depth over time, with time as the horizontal axis and anesthesia depth BIS as the vertical axis, output the trend curve of the change of the patient's anesthesia depth within a preset time period in the future.

5. A method for dynamic evaluation of anesthetic efficacy based on artificial intelligence, applied to the dynamic evaluation system for anesthetic efficacy based on artificial intelligence as described in any one of claims 1-4, characterized in that, Includes the following steps: Step 1: Collect the patient's real-time physiological parameters through medical monitoring equipment, construct a virtual anesthesia image based on the real-time physiological parameters, generate a virtual physiological model through pharmacokinetic algorithms, and output anesthetic efficacy assessment data; Step 2: Obtain anesthetic efficacy assessment data and the patient's actual physiological parameters during the operation. Compare the actual physiological parameters with the anesthetic efficacy assessment data, calculate the deviation value, adjust the virtual physiological model based on the reinforcement learning algorithm, and generate updated anesthetic efficacy assessment data based on the adjusted virtual physiological model. Step 3: Obtain updated anesthetic efficacy assessment data. Based on the adjusted virtual physiological model, predict the trend of changes in the patient's anesthesia depth within a preset time period. Combined with the preset conventional trend of anesthetic efficacy changes, analyze and determine the patient's physiological abnormalities. Simultaneously obtain the abnormal time corresponding to the physiological abnormality and the actual physiological parameters corresponding to the abnormal time. Perform retrospective analysis on the physiological abnormalities to determine the impact of anesthetic drugs on the physiological abnormalities. Step 4: Obtain the confirmed results, compare them with the preset safety thresholds for anesthetic efficacy, classify different warning levels based on the comparison results, generate warning information and send it to the physician's end.

Images