Drug delivery optimization method, device, equipment, medium and product

Abstract

The invention discloses a drug delivery optimization method and device, equipment, a medium and a product, and relates to the field of drug delivery, and the method comprises the steps: carrying out the feature extraction according to image data, and obtaining key features; the key features comprise drug concentration, time sequence change information and organization structure information; according to the key features, based on the drug concentration and release rate model, utilizing an artificial intelligence algorithm to predict the distribution condition of the drug in the body; optimizing drug delivery parameters by using a multi-objective optimization algorithm according to the distribution condition of the drugs in the body to obtain optimized drug delivery parameters; the optimized drug delivery parameters comprise dose, speed and delivery depth; carrying out delivery by utilizing the optimized drug delivery parameters and acquiring a real-time image after delivery; and performing feedback adjustment on the optimized drug delivery parameters by using the real-time image after delivery. The accuracy and safety of drug delivery can be improved.

Description

Technical Field

[0001] The present application relates to the field of drug delivery, and in particular to a drug delivery optimization method, device, equipment, medium and product. Background Art

[0002] Ankylosing spondylitis is a chronic inflammatory disease that primarily affects the spine and sacroiliac joints. Traditional treatments include systemic medications and local injections, but these approaches have significant limitations in terms of drug concentration, efficacy, and side effects. Systemic medications cannot achieve sufficient local drug concentrations, resulting in limited therapeutic efficacy, while local injections carry the risk of pain and infection, and it is difficult to precisely control drug release.

[0003] Therefore, a method that can improve the accuracy and safety of drug delivery is needed. Summary of the Invention

[0004] The purpose of this application is to provide a drug delivery optimization method, device, equipment, medium and product that can improve the accuracy and safety of drug delivery.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a method for optimizing drug delivery, comprising:

[0007] Acquiring image data; the image data includes a drug concentration distribution diagram, a time series dynamic image, and a tissue structure diagram;

[0008] Extracting features based on the image data to obtain key features; the key features include drug concentration, time series variation information, and tissue structure information;

[0009] Predicting the distribution of the drug in the body using an artificial intelligence algorithm based on the drug concentration and release rate model according to the key characteristics;

[0010] Optimizing drug delivery parameters using a multi-objective optimization algorithm according to the distribution of the drug in the body to obtain optimized drug delivery parameters; the optimized drug delivery parameters include dose, speed, and delivery depth;

[0011] delivering the drug using the optimized drug delivery parameters and acquiring real-time images after delivery;

[0012] The real-time images after the delivery are used to perform feedback adjustment on the optimized drug delivery parameters.

[0013] In a second aspect, the present application provides a drug delivery optimization device, comprising:

[0014] Dynamic imaging technology module, used to obtain image data; the image data includes drug concentration distribution diagram, time series dynamic image and tissue structure diagram;

[0015] A data acquisition and processing module is used to extract features based on the image data to obtain key features; the key features include drug concentration, time series change information and tissue structure information;

[0016] An artificial intelligence deduction algorithm module, configured to predict the distribution of the drug in the body using an artificial intelligence algorithm based on the key features and the drug concentration and release rate model;

[0017] a decision optimization module, configured to optimize drug delivery parameters using a multi-objective optimization algorithm according to the distribution of the drug in the body, thereby obtaining optimized drug delivery parameters; the optimized drug delivery parameters include dose, speed, and delivery depth;

[0018] an ultra-high-speed fluidics module, for delivering the drug using the optimized drug delivery parameters and acquiring real-time images after delivery;

[0019] A feedback and iteration module is used to perform feedback adjustment on the optimized drug delivery parameters using the real-time image after the delivery.

[0020] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the drug delivery optimization method.

[0021] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the drug delivery optimization method when executed by a processor.

[0022] In a fifth aspect, the present application provides a computer program product, comprising a computer program, which implements the drug delivery optimization method when executed by a processor.

[0023] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0024] The present application provides a drug delivery optimization method, device, equipment, medium and product, which predicts the distribution of the drug in the body by using an artificial intelligence algorithm based on the key characteristics and the drug concentration and release rate model; optimizes the parameters of drug delivery according to the distribution of the drug in the body using a multi-objective optimization algorithm to obtain optimized drug delivery parameters, which can improve the accuracy and safety of delivery, and uses the optimized drug delivery parameters to deliver the drug and obtain real-time images after delivery; uses the real-time images after delivery to feedback adjust the optimized drug delivery parameters, and after the drug delivery is performed, a feedback mechanism is also set up to further improve accuracy and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 A schematic diagram of a drug delivery optimization method;

[0027] Figure 2 A schematic diagram of a drug delivery optimization method;

[0028] Figure 3 A schematic diagram of the functional modules of a drug delivery optimization device;

[0029] Figure 4 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0031] In recent years, minimally invasive drug delivery technologies have garnered increasing attention. Ultrahigh-speed jet drug delivery systems, in particular, have attracted significant attention due to their ability to efficiently and accurately deliver drugs directly to the lesion. The key to this technology lies in precise control of jet velocity, drug dosage, and release depth to ensure that the drug reaches an effective concentration at the site of inflammation while minimizing damage to surrounding healthy tissue.

[0032] Currently, the existing drug delivery solutions mainly include the following:

[0033] 1. Ultrasound-guided drug delivery system:

[0034] Methods: Ultrasound imaging is used to monitor the distribution of drugs in tissues in real time. Doctors can adjust drug dosage and injection location based on real-time imaging data.

[0035] Steps: (1) Use ultrasound imaging to determine the location of the lesion, (2) perform local injection according to the preset drug dose and injection speed, (3) monitor the diffusion of the drug in the tissue and adjust the subsequent treatment plan based on the feedback information.

[0036] 2. Drug release system combining dynamic imaging and computational modeling:

[0037] Devices: These include dynamic imaging equipment (such as MRI or CT), drug delivery devices, and computational analysis platforms.

[0038] Structural composition: (1) Dynamic imaging equipment is used to obtain real-time images of drug release, (2) Drug delivery device is responsible for releasing drugs according to set parameters, and (3) Computational analysis platform receives imaging data, uses algorithms to analyze the diffusion dynamics of drugs in tissues, and predicts drug distribution.

[0039] 3. Minimally invasive drug delivery devices based on biocompatible materials:

[0040] Methods: Minimally invasive devices made of biocompatible materials are used to continuously release drugs to the target area through implantation or local injection.

[0041] The steps are as follows: (1) preparing a biocompatible material with drug loading capacity, (2) embedding the drug into the material and forming a structure such as microspheres or microneedles, (3) implanting or injecting the device into the sacroiliac joint area and utilizing the degradation characteristics of the material to achieve sustained release of the drug, and (4) monitoring the drug release effect and tissue response through regular imaging examinations.

[0042] Although the above technical solutions have made some progress in drug delivery, they still have the following limitations:

[0043] (1) Lack of individualized treatment: Existing methods generally use fixed drug dosages and release rates, which cannot be adjusted in real time based on individual patient differences, resulting in different treatment effects. This "one-size-fits-all" approach cannot effectively address the differences in disease conditions, tissue characteristics, and drug responses among different patients, resulting in unsatisfactory treatment results.

[0044] (2) Insufficient real-time feedback: Although some programs use dynamic imaging technology for monitoring, the evaluation is only performed after surgery or treatment, and there is a lack of real-time monitoring and feedback mechanism for the release dynamics during drug delivery. Existing drug delivery systems generally lack the ability to dynamically monitor drug release behavior and cannot obtain the distribution of drugs in tissues and release dynamics in real time. This lack of information makes it difficult for medical staff to adjust treatment plans to respond to emergencies during treatment, thereby affecting the safety and effectiveness of treatment.

[0045] (3) The prediction model is not accurate enough: Existing computational models are mostly based on general pharmacokinetic laws and fail to fully utilize individualized data and artificial intelligence technology to accurately predict the distribution and response of drugs in tissues.

[0046] (4) Inaccuracy in drug delivery: Existing drug delivery technologies, such as traditional systemic administration and local injection, often fail to achieve precise drug release in the lesion area. Systemic administration results in insufficient local drug concentration, while local injection carries the risk of pain and infection. This inaccuracy requires increasing the drug dosage to achieve therapeutic effects, thereby increasing the risk of side effects.

[0047] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0048] In an exemplary embodiment, Figure 1 As shown, a drug delivery optimization method is provided, which is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to a server as an example for description, including the following steps 101 to 106. Among them:

[0049] Step 101: Acquire image data; the image data includes a drug concentration distribution diagram, a time series dynamic image, and a tissue structure diagram;

[0050] Step 102: Extracting features based on the image data to obtain key features; the key features include drug concentration, time series variation information, and tissue structure information;

[0051] Step 103: Predicting the distribution of the drug in the body using an artificial intelligence algorithm based on the key features and the drug concentration and release rate model;

[0052] Step 104: Optimizing drug delivery parameters using a multi-objective optimization algorithm based on the distribution of the drug in the body to obtain optimized drug delivery parameters; the optimized drug delivery parameters include dose, speed, and delivery depth;

[0053] Step 105: delivering the drug using the optimized drug delivery parameters and acquiring a real-time image after delivery;

[0054] Step 106: Using the real-time image after delivery to perform feedback adjustment on the optimized drug delivery parameters.

[0055] Implement the above steps 101 to 106, predict the distribution of the drug in the body by using an artificial intelligence algorithm based on the key features and the drug concentration and release rate model; optimize the drug delivery parameters based on the distribution of the drug in the body using a multi-objective optimization algorithm to obtain optimized drug delivery parameters, which can improve the accuracy and safety of delivery, use the optimized drug delivery parameters to deliver the drug and obtain real-time images after delivery; use the real-time images after delivery to feedback adjust the optimized drug delivery parameters, and after drug delivery, set a feedback mechanism to further improve accuracy and safety.

[0056] In an exemplary embodiment, acquiring image data specifically includes: acquiring image data using an ultrasound imaging method or a magnetic resonance imaging method.

[0057] In an exemplary embodiment, feature extraction is performed based on the image data to obtain key features, specifically including: preprocessing the image data; determining the drug concentration in different tissues using a regional average method or a weighted average method for the preprocessed drug concentration distribution map; determining the time series transformation information of the preprocessed time series dynamic image using a dynamic curve fitting method, a time window analysis method, and a linear regression method; and performing image segmentation on the preprocessed tissue structure map to obtain tissue structure information.

[0058] In an exemplary embodiment, the process of constructing the drug concentration and release rate model specifically includes: organizing and structuring historical drug concentrations, historical temporal change information, and historical tissue structure information to obtain a time series with labels; performing time series analysis on the time series with labels to determine the rate and peak time of drug release; constructing a drug concentration and release rate model based on the distribution characteristics of the drug in the target tissue, the rate and peak time of drug release; the drug concentration and release rate model is a one-compartment model or a multi-compartment model.

[0059] In an exemplary embodiment, the artificial intelligence algorithm is a random forest model.

[0060] In an exemplary embodiment, the objective function of the multi-objective optimization algorithm is:

[0061] Minimize f(x) = ω1 (measurement error) + ω2 (speed error) + ω3 (delivery depth error)

[0062] Among them, f(x) is the objective function for quantifying the comprehensive error of drug delivery, representing the total error (dimensionless). The smaller the value, the more accurate the delivery. ω1, ω2, and ω3 are all weight coefficients, and x is the drug delivery parameter.

[0063] This application aims to address existing drug delivery systems, which suffer from uneven drug distribution within target tissues, difficulty in real-time monitoring of drug release kinetics, and insufficient individualized treatment. Specifically, traditional drug delivery methods are often limited by biological barriers, resulting in low drug distribution and absorption efficiency within the body, which in turn affects therapeutic efficacy. Furthermore, the lack of real-time monitoring and feedback mechanisms prevents timely adjustment of drug delivery parameters, leading to potential safety risks during treatment.

[0064] To this end, the present application combines dynamic imaging technology (such as ultrasound imaging and magnetic resonance imaging) to monitor the kinetic behavior of drug release in real time, and uses artificial intelligence big data deduction algorithms to accurately predict the distribution of drugs in tissues based on the individual differences of patients, specifically, accurately predict the distribution of drugs in sacroiliac joints and surrounding tissues. This method will effectively improve the accuracy and safety of drug delivery, ensure that potential problems can be identified in a timely manner and delivery parameters can be optimized, thereby achieving more efficient individualized treatment. Aiming at the individual differences of patients with ankylosing spondylitis, personalized and precise drug delivery is achieved. This solution not only optimizes the release kinetic behavior of drugs in the bone and joint area, but also improves the safety and effectiveness of treatment through real-time monitoring and intelligent decision-making. Compared with the existing technology solutions, the present application has obvious advantages in terms of individualized treatment, real-time feedback and the accuracy of prediction models, and has broad application prospects and commercial value.

[0065] The innovation of this application lies in:

[0066] (1) Real-time monitoring and dynamic adjustment: Through dynamic imaging technology, the release behavior of drugs can be monitored in real time. Drug delivery parameters can be adjusted based on immediate feedback to ensure optimal drug concentration at the lesion. By combining dynamic imaging technology, the release kinetics of drugs in human tissues can be monitored in real time, and drug distribution information can be obtained in a timely manner. This function enables medical staff to make adjustments based on real-time data, optimize drug delivery parameters, identify potential problems, and thus improve the safety and effectiveness of treatment.

[0067] (2) Individualized decision support: Artificial intelligence algorithms are used to analyze individual differences in patients and form individualized drug delivery decisions. This decision support system not only considers the patient's physiological characteristics, but also integrates historical treatment data, drug responses, and related biomarker information to make the drug delivery process more accurate and effective. Using artificial intelligence big data analysis algorithms, individualized drug delivery plans are designed based on the specific conditions of each patient (such as the degree of inflammation, pain intensity, etc.) to improve the targetedness and effectiveness of treatment.

[0068] (3) Optimization of the prediction model: This application uses big data analysis and machine learning algorithms to establish a prediction model for drug distribution in human tissues. This model can be updated in real time and continuously optimizes prediction accuracy based on patient feedback data, thereby improving treatment efficacy.

[0069] (4) Early identification of potential problems: Through real-time monitoring and data analysis of drug release dynamics, this application can identify possible side effects or adverse reactions in advance, thereby adjusting the treatment plan in a timely manner to ensure patient safety.

[0070] (5) Achieve precise drug delivery: By optimizing the rate and dosage of drug release, ensure that the drug can accurately reach the lesion area, increase local drug concentration, and reduce damage to healthy tissues, thereby achieving better therapeutic effects.

[0071] This application achieves the above goals through the following steps:

[0072] 1. Application of dynamic imaging technology: Using dynamic imaging technologies such as ultrasound imaging and magnetic resonance imaging, the distribution of drugs in human tissues can be monitored in real time. By acquiring high-resolution image data, the dynamic process of drug release can be understood.

[0073] 2. Data Acquisition and Processing: Collect and process data obtained during dynamic imaging. Use image analysis algorithms to extract the drug concentration distribution in different tissues and establish a drug concentration and release rate model.

[0074] 3. AI Big Data Deduction Algorithm: The processed data is fed into an AI Big Data Deduction Algorithm. Based on historical and real-time data, the algorithm analyzes individual patient characteristics and their impact on drug response, predicting drug distribution within human tissues.

[0075] 4. Development of personalized drug delivery plans: Based on the output results of the inference algorithm, a personalized drug delivery plan is tailored for each patient, including parameters such as dosage, release rate and administration method to suit the patient's specific needs.

[0076] 5. Real-time feedback and adjustment mechanism: During the drug delivery process, drug release is continuously monitored and real-time data is fed back to the system. If drug release does not meet expectations or an abnormal reaction occurs, the system can automatically adjust delivery parameters to ensure the safety and effectiveness of treatment.

[0077] Through these improvements, this application not only overcomes the shortcomings of existing technologies in drug delivery, such as inaccuracy and lack of personalized solutions, but also enables dynamic monitoring and real-time feedback, significantly improving the efficiency and safety of treatment. The implementation of this drug delivery decision-making method will provide a more scientific, reasonable, and effective solution for patient treatment, promoting the development of personalized medicine.

[0078] In practical applications, such as Figure 2 As shown, the method provided by this application includes the following steps.

[0079] Step 1: Dynamic imaging uses ultrasound imaging and magnetic resonance imaging to monitor the release and distribution of drugs in the human body in real time, acquiring relevant image data. The drug must first be delivered to the body via a safe needle-injection or needle-free method. This process aims to obtain pre-image data of the drug in the body, enabling subsequent real-time monitoring and analysis.

[0080] It is necessary to collect drug concentration distribution diagrams, time-series dynamic images, and tissue structure diagrams. These can be obtained through ultrasound imaging or magnetic resonance imaging technology. The specific methods are as follows:

[0081] (1) Drug concentration distribution diagram: shows the concentration distribution of drugs in different tissues, helping to analyze the release pattern of drugs.

[0082] Ultrasound imaging: After drug delivery, an ultrasound probe emits high-frequency sound waves to detect the echoes of the drug in tissues. Ultrasound imaging can provide real-time monitoring of drug concentrations. Combined with the acoustic properties of the drug (such as sound velocity and attenuation), the concentration distribution of the drug in different tissues can be calculated.

[0083] Magnetic resonance imaging (MRI): MRI technology is used in conjunction with contrast agents (such as certain metal ions or specific compounds) contained in the drug to generate images after drug injection. By monitoring the distribution of the contrast agent in the tissue, the drug concentration profile can be derived.

[0084] (2) Time series dynamic images: record the dynamic changes of drugs at specific time points and provide the time-concentration curve of drug release.

[0085] Dynamic ultrasound imaging: Ultrasound imaging is performed at different time points after drug delivery. By varying imaging parameters (such as acquisition frequency and image sequence), the dynamic changes of the drug within the tissue can be recorded. Dynamic images of the drug at specific time points can be generated, forming a time-concentration curve.

[0086] Dynamic MRI: Using a rapid acquisition sequence, multiple images are taken before and after drug injection to capture changes in drug distribution within tissues in real time. This method provides high-temporal resolution data, helping to study the dynamics of drug release.

[0087] (3) Tissue structure diagram: Combined with the use of contrast agents, clear images of tissue structure are obtained to assist in understanding the targeting of drugs.

[0088] Ultrasound imaging combined with contrast agents: When using ultrasound imaging, contrast agents (such as microbubbles) are injected. By enhancing the echogenicity of the contrast agents, clearer images of tissue structures can be obtained. This method can help identify the specific tissue targeted by the drug and assess its structural characteristics.

[0089] Dynamic MRI combined with contrast agents: When using dynamic MRI, the injection of specific contrast agents (such as gadolinium-based contrast agents) can significantly enhance tissue signals and provide high-resolution images of tissue structure. The distribution of the contrast agent can clearly identify the specific tissue targeted by the drug and assess its structural characteristics, helping to understand the behavior of the drug in the targeted tissue.

[0090] Step 2: Data acquisition and processing: organize and process the data collected by the dynamic imaging module to extract the distribution information and release kinetics characteristics of the drug in different tissues.

[0091] For the drug concentration distribution diagram, time series dynamic image, and organizational structure diagram collected in step 1, the following data collation and processing steps are required:

[0092] (1) Data preprocessing: De-noising, normalization, and enhancement are performed on the acquired image data to improve data quality and ensure the accuracy of subsequent analysis.

[0093] (2) Feature extraction: Use image processing technology to extract key features of drug concentration, temporal changes, and tissue structure.

[0094] The feature extraction process includes the following steps:

[0095] The drug concentration of the extracted drug is subjected to intensity analysis, that is, the intensity value of each pixel is analyzed and compared with the preset drug concentration standard to calculate the drug concentration in different tissues. The calculation can be performed using two methods: regional average or weighted average.

[0096] The regional average calculation method refers to averaging the pixel intensity values ​​within a specific area to estimate the drug concentration in that area. The calculation steps are as follows: (a) determining the region of interest of the drug concentration distribution map, that is, the pixel set containing the drug. Through image processing technology, the pixels with drug concentrations higher than the background noise are threshold-segmented to extract the pixel set containing the drug for determination; (b) calculating the sum of the intensity values ​​of all pixels in the area; and (c) dividing the sum by the number of pixels in the area. The calculation formula is:

[0097]

[0098] Among them C avg is the average concentration of the region, N is the total number of pixels in the region, I i is the intensity value of the i-th pixel.

[0099] For the weighted average method: it means giving different weights to different pixels during calculation, so as to more accurately reflect the distribution of the drug in the tissue. Usually, the weight can be set according to the intensity value or other characteristics of the pixel. Its calculation steps are: (a) Determine the region of interest of the drug concentration distribution map. (b) Assign a weight to each pixel. Usually, the intensity value of the pixel can be used as the weight. (c) Calculate the sum of the weighted intensities and divide it by the sum of the weights. The calculation formula is:

[0100]

[0101] Among them C w is the weighted average concentration, w i is the weight of the i-th pixel (can be set to I i The value of i is the intensity value of the i-th pixel, and N is the total number of pixels in the region.

[0102] It should be noted that the regional average method can be used when the pixel intensity values ​​in the region of interest are relatively uniform, or when performing preliminary analysis or quickly estimating drug concentrations. The weighted average method can be used when the pixel intensity values ​​within the region of interest vary significantly, or when improving the accuracy of the results is required, especially when analyzing the distribution of drugs in different tissues or cells.

[0103] To analyze the temporal changes of dynamic images, time-series image processing is required. This involves performing time series analysis on multiple frames of dynamic imaging data to extract a curve showing the drug concentration changes over time. This can be done by: (a) dynamic curve fitting: using polynomial regression or exponential decay models to fit the changes in drug concentration over time to obtain the drug release kinetics; and (b) time window analysis: setting a specific time window, calculating the drug concentration at each time point, and analyzing its changing trend to identify peaks and troughs in drug release.

[0104] For the above polynomial regression model, its basic form is:

[0105] C(t)=a n t n +a n-1 t n-1 +…+a1t+a0

[0106] Where: C(t) is the drug concentration at time t, a n 、a n-1 …a1 and a0 are coefficients to be determined, and n is the order of the polynomial, which can be selected according to the characteristics of the data.

[0107] The calculation steps are as follows: (1) Collect drug concentration data C(t i ), (2) select the appropriate polynomial order n, (3) use the least squares method to fit the data and solve for the coefficient a n 、a n-1 …a1, a0, (4) get the fitted concentration function C(t).

[0108] For the above exponential decay model, its basic form is:

[0109] C(t)=C0e -λt

[0110] Where: C(t) is the drug concentration at time t, C0 is the initial drug concentration, and λ is the decay constant, which represents the rate of drug release.

[0111] The calculation steps are as follows: (1) Collect drug concentration data C(t i ), (2) take logarithmic transformation: ln[C(t)] = ln(C0) - λt, (3) determine the coefficients C0 and λ through linear regression and establish the concentration function.

[0112] For the above-mentioned polynomial regression model and exponential decay model, if the drug concentration shows a nonlinear relationship with time, especially when the change trend is relatively smooth and can be effectively described by a polynomial, the polynomial regression model is used. If the drug concentration shows a monotonically decreasing trend with time, especially the elimination process of the drug in the body, such as drug metabolism and excretion, the exponential decay model is used.

[0113] For the above time window analysis, the basic process is:

[0114] (a) Setting the time window and data preparation: Determine the length of the time window Δt (e.g., 5 minutes) and collect drug concentration data C(t) at different time points.

[0115] (b) Calculate the average concentration within the window: Process the drug concentration data within each time window and calculate the average concentration of the window

[0116]

[0117] Where n is the number of time points in the window.

[0118] (c) Draw the concentration change curve: the average concentration of each time window Draw a curve to observe the change trend of drug concentration over time and identify peaks and troughs.

[0119] (d) Analyze results and adjust the plan: Based on the concentration change curve, evaluate the characteristics of drug release, identify potential therapeutic issues, and adjust subsequent drug delivery parameters to optimize the therapeutic effect.

[0120] Extracting these tissue structural features requires image segmentation. This involves using edge detection, region growing, or a watershed algorithm to segment the image and extract the outlines of the drug release zone and surrounding tissue. Specific methods include: (a) Edge detection: Using the Canny edge detection algorithm, the edges of the drug release zone are identified to help determine the distribution of the drug in different tissues. (b) Region growing: Starting from an initial seed point, the region is gradually expanded based on pixel intensity similarity to extract the drug release zone and the distribution of different tissue types.

[0121] Extracting tissue structural features requires calculating characteristic quantities. This involves quantitatively analyzing the extracted drug release regions and tissue types, and calculating key characteristic quantities, namely, tissue structural information. These include: (a) Area: The area of ​​the drug release region, reflecting the distribution of the drug within that region. (b) Shape characteristics: These characteristics, such as perimeter, circularity, and shape factor, characterize the shape of the drug release region. (c) Tissue type classification: By analyzing the intensity and distribution of tissue features, machine learning algorithms are used to classify tissue types and identify different biological tissues.

[0122] For the above calculation of key feature quantities, the basic process is as follows:

[0123] (a) Area calculation:

[0124] ① Binarize the drug release area of ​​the extracted tissue structure diagram to clearly separate the target area.

[0125] ② Use pixel counting method to calculate the total number of pixels in the target area of ​​the binary image,

[0126] ③ Convert to actual area based on pixel size: Area = number of pixels × actual area of ​​each pixel.

[0127] (b) Shape characteristics:

[0128] ① After calculating the contour of the drug release area, the perimeter P is calculated using the contour information.

[0129]

[0130] where d i is the distance between adjacent contour points.

[0131] ② Calculate the roundness (or shape factor), which is defined as:

[0132]

[0133] ③Record other shape features, such as shape factors, to help describe the geometric characteristics of the target area.

[0134] (c) Tissue type classification:

[0135] ① Extract the grayscale intensity and other features (such as texture features) of the tissue area to form a feature vector.

[0136] ② Use machine learning algorithms (such as random forest) to train feature vectors and establish a classification model.

[0137] ③ Input new data into the model, classify the tissue types, and output the identified different tissue types.

[0138] There is a close relationship between the area calculation, shape characteristics, and tissue type classification performed on the above-mentioned tissue structure diagram (area and shape characteristics jointly describe the geometric characteristics of the drug release area, and these geometric characteristics can be used as input features to help machine learning algorithms more accurately classify different tissue types, thereby optimizing drug delivery decisions). The processing of the three can be analyzed through a comprehensive processing flow.

[0139] (1) Feature relationship modeling: In this stage, a multidimensional feature space is constructed to associate the area and shape characteristics (such as perimeter, roundness, and shape factor) of the drug release region with the corresponding tissue type classification. By constructing a multidimensional feature space, a machine learning algorithm (such as random forest) is used to establish an association model between the area and shape characteristics of the drug release region and the tissue type classification. The machine learning algorithm is used to analyze the relationship between these features to identify important features that affect drug distribution and tissue type. This process can help discover the degree of influence of different tissue types on drug release and provide an important basis for subsequent decision-making.

[0140] (2) Data fusion and feature association analysis: The association model established in step (1) (feature relationship modeling) is integrated with dynamic imaging data to analyze the distribution of drugs in different tissues in real time. By using statistical analysis methods, the interaction between area, shape characteristics and tissue type is evaluated. This process will provide data support for the optimization of personalized drug delivery solutions, ensuring that the best drug delivery strategy can be found for the specific conditions of different patients.

[0141] Based on the above-mentioned use of image processing technology to extract key features of drug concentration, temporal changes and tissue type, these data can be converted into drug concentration and release rate models (such as one-compartment models or multi-compartment models) through the following steps to describe the distribution information and release kinetics of drugs in different tissues.

[0142] (1) Data organization and structuring: After extracting drug concentration, temporal changes, and tissue types, this information must first be organized and structured for subsequent analysis. This includes:

[0143] (This step is related to feature extraction from the three images in that data organization and structuring integrates the extracted drug concentration, temporal changes, and tissue types into an analyzable dataset, thereby providing the necessary basic information for establishing a drug concentration and release rate model.)

[0144] Drug concentration: Extract drug concentration data at different time points and in different tissues.

[0145] Time series changes: For the extracted drug concentration data, by recording the concentration values ​​at different time points, a complete time series is formed to help analyze the concentration distribution of drugs in different tissues.

[0146] Tissue type: Identify the characteristics of different tissues, such as blood vessels, muscle, or adipose tissue, to analyze the distribution of drugs in different media.

[0147] (2) Dynamic change analysis: Next, by analyzing the dynamic changes of the obtained concentration data, the kinetic characteristics of drug release can be obtained. This process usually includes:

[0148] Time series analysis: observe the changes in drug concentration over time and identify the rate and peak time of drug release.

[0149] Establish a mathematical model: Based on the temporal trend of concentration changes, select an appropriate drug concentration and release rate model (such as a one-compartment model or a multi-compartment model) and fit the model parameters. The mathematical model is constructed using the time-varying drug concentration data obtained through real-time dynamic imaging monitoring. By performing time-series analysis on this data and fitting a drug concentration and release rate model, the drug concentration and release rate in the target tissue are determined.

[0150] (3) Model parameter estimation and verification. Finally, the parameters in the drug concentration and release rate model are estimated by fitting the model to the extracted data. This includes:

[0151] Parameter fitting: Using nonlinear least squares or other optimization algorithms, adjust the parameters in the model (such as the elimination rate constant k) to minimize the error between the model predictions and the actual measurements.

[0152] Model validation: Verify the accuracy and reliability of the model through cross-validation or use of independent data sets to ensure that the established drug concentration and release rate model can accurately predict the distribution and release behavior of the drug in vivo.

[0153] Based on the above data conversion and post-processing, a drug concentration and release rate model (such as a one-compartment model or a multi-compartment model) is used to describe the distribution and elimination process of the drug in the body.

[0154] One-compartment model: Assuming that the drug is evenly distributed in the body, the drug concentration changes with time and can be described by a simple differential equation. The basic equation for the change of drug concentration with time is:

[0155] C(t)=C0e -kt

[0156] Where C(t) is the drug concentration at time t, C0 is the initial drug concentration, k is the elimination rate constant, and t is the time. The elimination rate constant k can be calculated by the following formula:

[0157]

[0158] Multicompartment models are used in pharmacokinetics to describe the complexity of drug distribution in the body, typically divided into two or more compartments. Take the common two-compartment model as an example: In this model, the drug in the body is considered to be distributed in two main "compartments": one is a rapidly distributing compartment (such as plasma) and the other is a slowly distributing compartment (such as tissue). The drug can be transferred between these two compartments, and each compartment has its own specific elimination rate. The basic equation of the two-compartment model is:

[0159]

[0160] Where C1 and C2 are the drug concentrations in chamber 1 and chamber 2, respectively, and k 12 is the transport rate constant from compartment 1 to compartment 2, k 21 is the transport rate constant from chamber 2 to chamber 1. The solution of the system usually needs to be solved using numerical methods, such as numerical integration methods such as the Euler method or the Runge-Kutta method.

[0161] The use scenarios of the one-compartment model include: (1) Simplified drug distribution: When the distribution of the drug in the body is relatively uniform and can almost be regarded as being distributed in a whole "chamber", the one-compartment model is suitable. This model assumes that the drug is evenly distributed in the entire body fluid volume within a given time. (2) Short-term drug action: If the drug's action time is short and does not involve a complex distribution process (such as the drug quickly enters the blood and is quickly eliminated), the one-compartment model can effectively describe the change of drug concentration over time. (3) Preliminary research or screening: In the early stages of drug development, a one-compartment model is usually used for preliminary evaluation in order to quickly obtain the basic kinetic characteristics of the drug. The use scenarios of the multi-compartment model: (1) Complex drug distribution: When the drug is distributed at different speeds and in different ways in different tissues or organs, the multi-compartment model is suitable. The multi-compartment model can better describe the dynamic distribution of the drug in different parts, taking into account the transport between tissues and the different elimination rates of the drug. (2) Long-term drug action: For treatment plans that require long-term administration, or when the drug has a long half-life in the body, the multi-compartment model can more accurately reflect the change of drug concentration over time. (3) Interaction between drugs and biological tissues: When the distribution and elimination processes of drugs are significantly affected by tissue properties (such as blood flow, tissue density, cell absorption capacity, etc.), multi-compartment models can more accurately capture this complex interaction.

[0162] Step 3: Artificial intelligence deduction algorithm, based on the processed data, uses artificial intelligence algorithm to analyze the individual differences of patients, predict the distribution of drugs in the body, and identify potential risk points.

[0163] Based on the data processing completed in step 2, the established drug concentration and release rate model is applied to the artificial intelligence algorithm. Here, the random forest model is used. This model is suitable for analyzing individual differences in patients and predicting the distribution of drugs in the body. The use process of the random forest model is as follows:

[0164] (1) Data preparation and feature selection

[0165] Collect and organize relevant drug release kinetics data and patient-specific information. Clean the data to remove missing and outliers to ensure data quality. Select appropriate feature variables, which may include drug concentration, injection velocity, patient physiological characteristics, and other relevant clinical data.

[0166] (2) Model training

[0167] Split the prepared dataset into a training set and a test set. Typically, the training set comprises 70%-80% of the total data. Use the training set to train a random forest model, determining the number of trees (often requiring cross-validation to determine the optimal value) and other hyperparameters (such as maximum depth and minimum number of sample splits). The random forest model learns from multiple decision trees, obtaining predictions from each tree and generating the final model output through a voting mechanism.

[0168] (3) Model prediction and evaluation

[0169] Use the trained model to predict the test set and obtain the distribution of the drug in the patient's body. Evaluate the model's predictive performance using metrics such as accuracy, recall, and F1-score to ensure the model's effectiveness and reliability.

[0170] (4) Risk identification and decision support

[0171] Based on the model's predictions, potential risk points, such as uneven drug distribution or overdose, are identified. Combining clinical feedback with dynamic imaging data, the drug delivery plan is adjusted in real time to ensure the safety and effectiveness of treatment.

[0172] The usage formula of the random forest model can be expressed as:

[0173]

[0174] in is the final prediction result, which represents the model's predicted value for the input feature X. T is the number of trees in the random forest, which is usually determined by hyperparameter tuning. t (X) is the prediction result of the t-th decision tree for the input feature X.

[0175] By constructing multiple decision trees and combining their predictions, the random forest model effectively reduces the risk of overfitting and improves prediction accuracy and stability. In specific applications, this formula can be used to calculate individualized drug distribution predictions for each patient, providing a scientific basis for drug delivery decisions.

[0176] In step 2, a drug concentration and release rate model was established. The core of this model is the distribution data and release characteristics of the drug in different tissues collected through dynamic imaging technology. The model extracts the drug's behavior pattern in the body, including important parameters such as release rate and peak time. In step 3, a random forest model was used. The construction of this model relies on the drug concentration and release rate kinetic characteristic data obtained in step 2. Specifically, the drug concentration and release rate model in step 2 provides the following basic support for the random forest model in step 3:

[0177] (1) Basic data: The data collected and processed in step 2 provide the necessary foundation for the input of the random forest model. These data include changes in drug concentration in different tissues, release timing information, etc., which are the key to making individualized predictions.

[0178] (2) Kinetic parameters: The drug concentration and release rate kinetic characteristics extracted in step 2, such as the initial concentration, distribution velocity, and duration of the drug, are the basis for the random forest model to analyze individual differences in patients. These parameters enable the model to understand the behavior of the drug in the body, thereby improving the accuracy of predictions.

[0179] (3) Feature Engineering: The model in step 2 not only generates raw data but also extracts features related to drug distribution through data processing and analysis. These features provide more refined inputs for the random forest model, helping the model better capture individual differences between patients.

[0180] Therefore, the drug concentration and release rate model established in step 2 provides a comprehensive data basis and parameter support for the random forest model, enabling the random forest model to more accurately analyze individual differences among patients, predict the distribution of drugs in the body, and identify potential risk points. This hierarchical relationship ensures the effectiveness and reliability of the method of this application in making personalized drug delivery decisions.

[0181] Step 4: Decision optimization: Based on the prediction results of the artificial intelligence module, optimize the parameters of drug delivery (such as dosage, speed and delivery depth) and formulate a personalized treatment plan.

[0182] Based on the prediction results in step 3, the following steps are needed to optimize drug delivery parameters (such as dose, rate, and delivery depth) and develop a personalized treatment plan. The specific steps are as follows:

[0183] (1) Data analysis and demand identification

[0184] First, we collect patient-specific characteristics and drug distribution predictions derived through dynamic imaging technology and artificial intelligence algorithms. This data will be used to identify patients' specific needs, such as drug concentration, distribution, and physiological response in target tissues.

[0185] (2) Establishment of parameter optimization model

[0186] An optimization model is established based on the individual characteristics of the patient and the predicted results. The model can use a multi-objective optimization algorithm, and its objective function can be expressed as:

[0187] Minimize f(x) = ω1 (measurement error) + ω2 (speed error) + ω3 (delivery depth error)

[0188] Where x is the parameter of drug delivery (dose, rate, delivery depth), ω1, ω2, and ω3 are weight coefficients reflecting the importance of each target.

[0189] (3) Parameter optimization process

[0190] Initial parameter setting: Initial drug delivery parameters are set based on historical data and clinical experience. Prediction results are then fed into the optimization model, including the output of the AI ​​module (e.g., predicted drug distribution within tissues). Through iteration, the drug dosage, delivery rate, and depth are adjusted, the objective function f(x) is calculated, and the optimization algorithm is used to find the optimal solution until the objective function reaches its minimum or meets the specified convergence criteria.

[0191] (4) Development of personalized treatment plans

[0192] Based on the optimization results, a personalized drug delivery plan is generated, including specific parameters such as dosage, release rate, and administration method. This plan must fully consider the physiological characteristics and pathological conditions of each patient to ensure the effectiveness and safety of the treatment.

[0193] Step 5: Ultra-high-speed jet: Generate an ultra-high-speed jet based on optimized parameters to accurately deliver the drug to the target tissue, ensuring the effective release of the drug in the lesion area.

[0194] Based on the parameters optimized in step 4, an ultra-high-speed jet is generated to ensure that the drug can be accurately delivered to the target tissue and effectively released in the lesion area. The specific process is as follows:

[0195] (1) Parameter setting and optimization

[0196] ① Determine key parameters: Based on the optimization results of step 4, select appropriate jet velocity (V), dose (D), and delivery depth (H). These parameters are the basis for achieving precise delivery.

[0197] ②Formula application: Use the dynamic pressure formula to calculate the required jet velocity to ensure that the drug can effectively penetrate the biological barrier and reach the target tissue.

[0198]

[0199] Where P is the dynamic pressure (Pa), ρ is the density of the drug solution (kg / m 3 ), v is the jet velocity (m / s).

[0200] (2) Generation of ultra-high-speed jets

[0201] ① Equipment adjustment: Use an ultra-high-speed jet needle-free syringe and adjust the equipment parameters (such as voltage, current, pulse frequency, etc.) to make the jet velocity reach the predetermined value (V) to generate a high-energy drug jet.

[0202] ②Drug dosage control: Based on the optimized dosage (D), ensure that the amount of drug injected each time meets individual needs and that the drug concentration in the target area reaches an effective level.

[0203] (3) Real-time monitoring of drug delivery process

[0204] ① Dynamic imaging technology: During the drug delivery process, combined with ultrasound imaging or magnetic resonance imaging technology, the distribution and diffusion of drugs in tissues can be monitored in real time to obtain the kinetic data of drug release.

[0205] ②Data analysis and feedback: Analyze monitoring data in real time, evaluate the concentration distribution of the drug in the target tissue, and adjust the jet parameters based on the results, such as increasing the jet speed or dose, to ensure that the drug can effectively reach the lesion area.

[0206] (4) Optimization and adjustment

[0207] ① Iterative Update: Continuously optimize the drug delivery process based on real-time monitoring and feedback data. Dynamically adjust jet parameters. If the drug fails to effectively reach the lesion area, the jet velocity (V) and dose (D) can be adjusted in real time to ensure better penetration of the target tissue. Model Update: Based on artificial intelligence algorithms, the drug delivery model is iteratively updated using acquired dynamic data to generate more accurate predictions to meet individual patient needs.

[0208] ② Calculation of drug release efficiency: Based on the monitoring data and the concentration distribution of the drug in the tissue, the following formula is used to evaluate the efficiency of drug release (E):

[0209]

[0210] Among them C target is the actual concentration of the drug in the target tissue (mg / mL), C dose is the desired drug concentration after injection (mg / mL).

[0211] Through the above four processes, step five can realize the process of generating ultra-high-speed jet based on optimized parameters, ensuring that the drug can be accurately delivered to the target tissue and effectively released in the lesion area.

[0212] Step 6: Feedback and iteration mechanism: continuously collect real-time data during the treatment process, evaluate the effect of drug delivery, and feed it back to the artificial intelligence module for iterative optimization to improve decision-making accuracy and treatment effect.

[0213] Based on the above five steps, it is necessary not only to continuously collect real-time data and evaluate the drug delivery effect, but also to feed back to the artificial intelligence module for iterative optimization.

[0214] The specific steps for continuously collecting real-time data during treatment and evaluating the effectiveness of drug delivery are:

[0215] ① Dynamic data acquisition: Utilizing dynamic imaging techniques (such as ultrasound and magnetic resonance imaging), the distribution of drugs in target tissues is monitored in real time. These techniques capture data such as drug concentration, distribution range, and release rate. Changes in drug concentration at different time points and in different tissues are recorded to facilitate subsequent analysis.

[0216] ② Data processing and standardization: Preprocess the collected real-time data, including data cleaning, denoising, and standardization, to ensure consistency and comparability. This generates a structured data set to facilitate subsequent analysis and modeling.

[0217] ③ Setting of effect evaluation indicators: Determine the key effect evaluation indicators of drug delivery, such as the drug in the target area

[0218] Effective drug release rate E:

[0219]

[0220] ④ Real-time effect evaluation: Analyze real-time data based on set evaluation indicators to evaluate the effect of drug delivery, including the consistency and effectiveness of drug distribution. Generate an effect evaluation report to provide a reference for subsequent decision-making.

[0221] Feedback is sent to the artificial intelligence module for iterative optimization. The specific steps are as follows:

[0222] ① Data Collection and Preprocessing: During drug delivery, dynamic imaging techniques (such as ultrasound and magnetic resonance imaging) are used to monitor drug release kinetics in real time. This includes key metrics such as drug concentration, distribution, and release rate. The collected raw data is cleaned, standardized, and de-noised to ensure data quality, generating a high-quality dataset for subsequent analysis.

[0223] ② Data Analysis and Model Update: The data analysis module conducts in-depth analysis of preprocessed data to identify key factors affecting drug release and therapeutic efficacy. For example, this module analyzes the impact of individual patient differences on drug response. The analysis results are input into the artificial intelligence module, which updates the machine learning model based on the latest data. This process can be evaluated using the following formula to evaluate the accuracy of the model's predictions:

[0224]

[0225] Among them, true positive (TP): when the model predicts that the drug concentration is within the effective range and the actual observed concentration is also within the effective range, it is counted as a true positive; true negative (TN): when the model predicts that the drug concentration is not within the effective range and the actual observed concentration is also not within the effective range, it is counted as a true negative; false positive (FP): when the model predicts that the drug concentration is within the effective range, but the actual observed concentration is not within the effective range, it is counted as a false positive; false negative (FN): when the model predicts that the drug concentration is not within the effective range, but the actual observed concentration is within the effective range, it is counted as a false negative.

[0226] ③ Real-time prediction and risk assessment: At each stage of drug delivery, real-time data collected is fed into an updated model to generate real-time predictions of drug distribution and concentration in different tissues. This provides real-time support for treatment planning. Based on these real-time predictions, risk assessments are performed to identify potential side effects or adverse reactions. This assessment can help adjust treatment plans to ensure patient safety.

[0227] ④ Decision optimization and feedback loop: Based on the risk assessment results and real-time predictions of drug distribution, adjust the parameters of drug delivery (such as dosage, release rate, and administration method) and formulate a personalized treatment plan. Feedback the optimized decision and parameter settings to the drug delivery system to adjust the subsequent drug delivery process. At the same time, ensure that new real-time data is continuously collected and re-input into the artificial intelligence module to form a closed-loop feedback mechanism. In this way, the entire system can continuously learn and adapt, thereby improving the safety and effectiveness of drug delivery.

[0228] Based on the same inventive concept, the present application also provides a drug delivery optimization device for implementing the aforementioned drug delivery optimization method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more drug delivery optimization device embodiments provided below can be found in the above-mentioned limitations of the drug delivery optimization method and will not be repeated here.

[0229] In an exemplary embodiment, Figure 3 As shown, a drug delivery optimization device is provided, comprising:

[0230] Dynamic imaging technology module 301 is used to obtain image data; the image data includes drug concentration distribution diagram, time series dynamic image and tissue structure diagram;

[0231] The data acquisition and processing module 302 is used to extract features based on the image data to obtain key features; the key features include drug concentration, time series variation information and tissue structure information;

[0232] An artificial intelligence deduction algorithm module 303 is used to predict the distribution of the drug in the body using an artificial intelligence algorithm based on the key features and the drug concentration and release rate model;

[0233] A decision optimization module 304 is configured to optimize drug delivery parameters using a multi-objective optimization algorithm based on the distribution of the drug in the body to obtain optimized drug delivery parameters; the optimized drug delivery parameters include dose, speed, and delivery depth;

[0234] An ultra-high-speed jet module 305 is used to deliver the drug using the optimized drug delivery parameters and obtain real-time images after delivery;

[0235] The feedback and iteration module 306 is configured to perform feedback adjustment on the optimized drug delivery parameters using the real-time images after the delivery.

[0236] In an exemplary embodiment, Figure 4As shown, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the drug delivery optimization method.

[0237] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the above-mentioned method embodiments when executed by a processor.

[0238] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the above method embodiments are implemented.

[0239] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0240] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A drug delivery optimization method, characterized in that: The drug delivery optimization method comprises: Acquiring image data; the image data includes a drug concentration distribution diagram, a time series dynamic image, and a tissue structure diagram; Extracting features based on the image data to obtain key features; the key features include drug concentration, time series variation information, and tissue structure information; Predicting the distribution of the drug in the body using an artificial intelligence algorithm based on the drug concentration and release rate model according to the key characteristics; Optimizing drug delivery parameters using a multi-objective optimization algorithm according to the distribution of the drug in the body to obtain optimized drug delivery parameters; the optimized drug delivery parameters include dose, speed, and delivery depth; delivering the drug using the optimized drug delivery parameters and acquiring real-time images after delivery; The real-time images after the delivery are used to perform feedback adjustment on the optimized drug delivery parameters.

2. The drug delivery optimization method according to claim 1, characterized in that Obtain image data, including: Image data is acquired using an ultrasound imaging method or a magnetic resonance imaging method.

3. The drug delivery optimization method according to claim 1, characterized in that Feature extraction is performed based on the image data to obtain key features, specifically including: Preprocessing the image data; Determining the drug concentrations in different tissues using a regional average method or a weighted average method for the drug concentration distribution map after pretreatment; The pre-processed time series dynamic image is used to determine the time series transformation information by using the dynamic curve fitting method, the time window analysis method and the linear regression method; The preprocessed tissue structure chart is subjected to image segmentation to obtain tissue structure information.

4. The drug delivery optimization method according to claim 1, characterized in that The process of constructing the drug concentration and release rate model specifically includes: According to the historical drug concentration, historical time series change information and historical organizational structure information, it is sorted and structured to obtain a time series with labels; Performing a time series analysis on the labeled time series to determine the rate and peak time of drug release; A drug concentration and release rate model is constructed based on the distribution characteristics of the drug in the target tissue, the rate of drug release and the peak time; the drug concentration and release rate model is a one-compartment model or a multi-compartment model.

5. The drug delivery optimization method according to claim 1, characterized in that The artificial intelligence algorithm is a random forest model.

6. The drug delivery optimization method according to claim 1, characterized in that The objective function of the multi-objective optimization algorithm is: Minimize f(x) = ω1·(metering error) + ω2·(velocity error) + ω3·(delivery depth error) where f(x) is the objective function for quantifying the comprehensive error of drug delivery, ω1, ω2, and ω3 are weight coefficients, and x is the drug delivery parameter.

7. A drug delivery optimization device, characterized in that The drug delivery optimization device comprises: Dynamic imaging technology module, used to obtain image data; the image data includes drug concentration distribution diagram, time series dynamic image and tissue structure diagram; A data acquisition and processing module is used to extract features based on the image data to obtain key features; the key features include drug concentration, time series change information and tissue structure information; An artificial intelligence deduction algorithm module, configured to predict the distribution of the drug in the body using an artificial intelligence algorithm based on the key features and the drug concentration and release rate model; a decision optimization module, configured to optimize drug delivery parameters using a multi-objective optimization algorithm according to the distribution of the drug in the body, thereby obtaining optimized drug delivery parameters; the optimized drug delivery parameters include dose, speed, and delivery depth; an ultra-high-speed fluidics module, for delivering the drug using the optimized drug delivery parameters and acquiring real-time images after delivery; A feedback and iteration module is used to perform feedback adjustment on the optimized drug delivery parameters using the real-time image after the delivery.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the drug delivery optimization method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the drug delivery optimization method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the drug delivery optimization method according to any one of claims 1 to 6 is implemented.

Images