CN-121983228-A - Accurate regulation and control method and system for post-operation medication of bladder based on artificial intelligence

CN121983228ACN 121983228 ACN121983228 ACN 121983228ACN-121983228-A

Abstract

The invention provides an accurate regulation and control method and system for post-operation medication of bladder based on artificial intelligence, which relate to the technical field of artificial intelligence and comprise the steps of performing feature extraction and semantic mapping based on a multi-mode feature fusion technology by acquiring post-operation physiological index data, medication history data and clinical symptom data of a patient, constructing individual state characterization of the patient, calculating pharmacokinetic parameters and curative effect response curves under different medication schemes through time sequence dependency modeling, generating a candidate medication regulation and control scheme set by combining a medication interaction constraint rule, solving an optimal medication regulation and control scheme under the constraint conditions of maximized curative effect, minimized side effect and medication cost, dynamically adjusting and optimizing a target weight coefficient according to the risk feature of the patient, acquiring actual curative effect data and adverse reaction data after medication adjustment, forming closed loop feedback, and realizing continuous iterative optimization of drug effect prediction parameters and weight coefficients.

Inventors

LIU XIAOXUAN
HOU CUICUI
ZHANG MUWEI
YANG CHUNYU

Assignees

中国人民解放军总医院第一医学中心

Dates

Publication Date: 20260505
Application Date: 20260311

Claims (9)

1. The accurate regulation and control method for the post-operation medication of the bladder based on artificial intelligence is characterized by comprising the following steps: The method comprises the steps of acquiring physiological index data, medication history data and clinical symptom data of a patient after operation, carrying out feature extraction and semantic mapping on the physiological index data, the medication history data and the clinical symptom data based on multi-mode feature fusion to obtain individualized state representation of the patient; According to the individuation state representation of the patient, calculating pharmacokinetic parameters and curative effect response curves under different medication schemes through time sequence dependency modeling, and generating a candidate medication regulation scheme set by combining with a medication interaction constraint rule; Solving an optimal medication regulation scheme under the constraint conditions of maximizing curative effect, minimizing side effect and medication cost according to the candidate medication regulation scheme set and postoperative recovery targets, and dynamically adjusting weight coefficients of all optimization targets based on risk characteristics in the personalized state representation of the patient; and executing medication adjustment based on the optimal medication adjustment scheme, collecting postoperative actual curative effect data and adverse reaction data to form feedback information, and iteratively updating the drug effect prediction parameters and the weight coefficients by using the feedback information.
2. The method of claim 1, wherein performing feature extraction and semantic mapping on the physiological index data, the medication history data, and the clinical symptom data based on multi-modal feature fusion, obtaining a patient personalized state characterization comprises: Extracting time sequence features of the physiological index data to obtain a physiological index time sequence feature vector, carrying out medicine effect association analysis on the medicine history data to obtain a medicine history association feature vector, and carrying out symptom severity quantification on the clinical symptom data to obtain a clinical symptom quantification feature vector; And performing cross-modal semantic alignment on the physiological index time sequence feature vector, the medication history association feature vector and the clinical symptom quantification feature vector, calculating association weights among modal features through an attention weighting mechanism, performing self-adaptive fusion on the modal features based on the association weights to obtain a fusion feature vector, and performing nonlinear mapping transformation on the fusion feature vector to obtain the individualized state representation of the patient.
3. The method of claim 1, wherein calculating pharmacokinetic parameters and efficacy response curves for different drug regimens by time-series dependency modeling based on the patient individualization state characterization, and generating a set of candidate drug administration regulatory regimens in combination with drug interaction constraint rules comprises: Constructing a patient-specific drug effect prediction reference state based on patient individuation state representation, calculating a residual influence coefficient of the prior drug on the current drug effect through recursive time sequence dependency modeling according to the drug effect prediction reference state and historical drug time sequence data, and obtaining a time sequence residual influence matrix; Coupling operation is carried out on the time sequence residual influence matrix and a drug solution to be evaluated, and a pharmacokinetic parameter change track and a curative effect response curve under the drug solution to be evaluated are predicted; acquiring a drug interaction constraint rule base, and carrying out security constraint screening on the curative effect response curve according to the drug interaction constraint rule base to remove a drug use scheme violating the drug interaction constraint; And combining and arranging the drug administration schemes screened by the safety constraint to generate a candidate drug administration regulation scheme set.
4. The method of claim 3, wherein coupling the time-series residual influence matrix to the drug regimen to be evaluated, and predicting the pharmacokinetic parameter profile and efficacy response curve for the drug regimen to be evaluated comprises: converting the information of the drug dosage and the drug administration time in the drug proposal to be evaluated into a time sequence input vector, and carrying out time-by-time weighted coupling operation on the time sequence input vector and a time sequence residual influence matrix to obtain a corrected metabolic parameter sequence containing residual influence; Calculating the drug absorption rate, the distribution rate and the elimination rate at corresponding moments based on the metabolic parameter values at each moment in the modified metabolic parameter sequence, and combining the drug absorption rate, the distribution rate and the elimination rate at continuous moments to form a pharmacokinetic parameter change track; According to the drug absorption rate and the elimination rate at each moment in the pharmacokinetic parameter variation track, calculating the drug blood concentration value at each moment, arranging the drug blood concentration values at each moment according to time sequence to form a blood concentration time sequence curve, and mapping the blood concentration time sequence curve to a curative effect space to obtain a curative effect response curve.
5. The method of claim 1, wherein solving for the optimal medication control scheme under the constraints of maximizing efficacy, minimizing side effects, and cost of medication based on the set of candidate medication control schemes and a post-operative recovery goal comprises: identifying a patient risk level according to risk characteristics in the patient individuation state representation, and respectively endowing an adaptive weight coefficient to a curative effect maximization target, a side effect minimization target and a medication cost constraint target based on the patient risk level to obtain a patient exclusive target weight configuration; Carrying out weighted summation operation on the patient-specific target weight configuration and the curative effect predicted value, the side effect predicted value and the medication cost value of each candidate solution in the candidate medication regulation and control solution set to obtain the comprehensive evaluation score of each candidate solution; And sorting the candidate schemes in the candidate medication regulation scheme set according to the comprehensive evaluation score, and selecting the candidate scheme with the highest comprehensive evaluation score from the sorting result as the optimal medication regulation scheme.
6. The method of claim 1, wherein performing medication adjustments based on the optimal medication adjustment scheme and collecting post-operative actual efficacy data and adverse reaction data to form feedback information comprises: generating a medication execution instruction according to the medicine type, the medication dosage and the medication time information in the optimal medication regulation scheme, sending the medication execution instruction to a medication execution terminal to finish medication adjustment, and recording an execution time stamp and an execution state identifier of the medication execution instruction; In a preset monitoring period after the medication execution instruction is executed, acquiring physiological index change data of a patient as postoperative actual curative effect data through physiological monitoring equipment, and acquiring adverse reaction symptom description and severity identification of the patient as adverse reaction data through an adverse reaction acquisition interface; And carrying out association binding on the execution time stamp, the execution state identifier, the postoperative actual curative effect data and the adverse reaction data to generate feedback information containing medication execution information and actual response information.
7. An artificial intelligence based accurate post-operative drug administration control system for implementing the method of any one of the preceding claims 1-6, comprising: The system comprises a data acquisition unit, a data processing unit and a data processing unit, wherein the data acquisition unit is used for acquiring physiological index data, medication history data and clinical symptom data of a patient after operation; the scheme generating unit is used for calculating pharmacokinetic parameters and curative effect response curves under different medication schemes through time sequence dependency modeling according to the personalized state representation of the patient, and generating a candidate medication regulation scheme set by combining with a medication interaction constraint rule; The scheme optimizing unit is used for solving an optimal medication regulating scheme under the constraint conditions of maximizing curative effect, minimizing side effect and medication cost according to the candidate medication regulating scheme set and postoperative recovery targets, and dynamically adjusting the weight coefficient of each optimizing target based on the risk characteristics in the personalized state representation of the patient; And the feedback updating unit is used for executing medication adjustment based on the optimal medication regulation scheme, collecting postoperative actual curative effect data and adverse reaction data to form feedback information, and iteratively updating the drug effect prediction parameters and the weight coefficients by utilizing the feedback information.
8. An electronic device, comprising: A processor; A memory for storing processor-executable instructions; Wherein the processor is configured to invoke the instructions stored in the memory to perform the method of any of claims 1 to 6.
9. A computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method of any of claims 1 to 6.

Description

Accurate regulation and control method and system for post-operation medication of bladder based on artificial intelligence Technical Field The invention relates to an artificial intelligence technology, in particular to an accurate regulation and control method and system for post-operation medication of a bladder based on artificial intelligence. Background Bladder surgery is a common surgical technique in urology, and post-operation medication management directly affects the recovery quality and the complication rate of patients. Traditional post-operative bladder medication schemes mainly depend on experience judgment of clinicians, and standard medication doses and frequencies are formulated by combining basic physiological indexes of patients and operative types. In clinical practice, a doctor usually refers to a recommended dosage range of a drug instruction, carries out linear adjustment according to basic parameters such as age, weight, liver and kidney functions of a patient, and judges the drug effect according to postoperative conventional monitoring indexes such as body temperature, blood pressure, urine volume and the like. This empirically driven mode of administration can meet basic therapeutic needs in most routine cases, creating a more mature clinical pathway system. Because the individual differences of patients are obvious, the metabolic rate, the peak time of the blood concentration and the tissue distribution characteristic difference of the same dosage of the medicine in different patients can be several times, so that the phenomenon of insufficient medication or excessive medication of partial patients occurs. The traditional scheme is difficult to dynamically capture the continuous change of the physiological state of a patient in the postoperative recovery process, the medication parameters cannot be timely adjusted according to the real-time condition fluctuation, and the problems of medication lag or improper adjustment frequency often occur. When multiple drugs are used in combination, the existing scheme lacks a systematic evaluation mechanism for complex drug interactions, and can cause unexpected effects of enhanced or weakened drug efficacy, and increase the risk of adverse reactions. In the aspect of efficacy evaluation, the comprehensive benefit of a drug administration scheme is difficult to accurately quantify simply by relying on subjective symptom description and limited objective indexes, and the accurate balance among efficacy, safety and economy cannot be realized, so that the overall treatment efficiency is affected. Disclosure of Invention The embodiment of the invention provides an accurate bladder postoperative drug regulation and control method and system based on artificial intelligence, which can solve the problems in the prior art. In a first aspect of an embodiment of the present invention, there is provided an artificial intelligence-based accurate post-operative drug administration control method for bladder, including: Performing feature extraction and semantic mapping on the physiological index data, the medication history data and the clinical symptom data based on multi-mode feature fusion to obtain individualized state characterization of the patient; According to the individuation state representation of the patient, calculating pharmacokinetic parameters and curative effect response curves under different medication schemes through time sequence dependency modeling, and generating a candidate medication regulation scheme set by combining with a medication interaction constraint rule; Solving an optimal medication regulation scheme under the constraint conditions of maximizing curative effect, minimizing side effect and medication cost according to the candidate medication regulation scheme set and postoperative recovery targets, and dynamically adjusting weight coefficients of all optimization targets based on risk characteristics in the personalized state representation of the patient; and executing medication adjustment based on the optimal medication adjustment scheme, collecting postoperative actual curative effect data and adverse reaction data to form feedback information, and iteratively updating the drug effect prediction parameters and the weight coefficients by using the feedback information. Performing feature extraction and semantic mapping on the physiological index data, the medication history data and the clinical symptom data based on multi-modal feature fusion, and obtaining the individualized state representation of the patient comprises the following steps: Extracting time sequence features of the physiological index data to obtain a physiological index time sequence feature vector, carrying out medicine effect association analysis on the medicine history data to obtain a medicine history association feature vector, and carrying out symptom severity quantification on the clinical symptom data to obtain a clinical symptom quantif