Angiography high-pressure injection system

Abstract

The present invention belongs to the field of medical imaging technology. The present invention discloses a high-pressure injection system for angiography, comprising: obtaining patient physiological data information, and calculating individualized contrast agent injection parameter coefficients based thereon; obtaining target patient vascular area anatomical structure information, and performing pressure assessment based on hemodynamic characteristics to obtain vascular pressure bearing capacity and injection pressure adjustment factors; performing physiological state assessment based on patient physiological data information, obtaining physiological state influencing factors and generating risk marker values; performing multi-source data fusion according to the injection pressure adjustment factors, physiological state risk marker values ​​and historical angiography records to construct angiography safety degree for a vascular sub-area; obtaining an optimal injection parameter path based on angiography safety degree combined with a deep reinforcement learning algorithm, and controlling a high-pressure injection device to perform an injection operation; the present invention realizes precise and personalized control of contrast agent injection, significantly improving the safety and imaging effect of angiography examinations.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and more particularly to a high-pressure injection system for angiography. Background Art

[0002] With the widespread application of medical imaging technology in clinical diagnosis, contrast agent injection has become a key link in improving image quality and diagnostic accuracy. However, existing contrast agent injection systems have the following technical defects:

[0003] Existing contrast agent injection systems have limited patient data collection capabilities, relying primarily on discrete medical records or single parameter monitoring. They lack a systematic, standardized data integration mechanism and are unable to fully capture key information such as a patient's physiological characteristics, renal function status, and historical contrast response, making it difficult to develop personalized injection plans.

[0004] Traditional injection systems generally lack precise parameter calculation mechanisms and often use a weight-based linear dose adjustment method. This method fails to fully consider multidimensional factors such as patient age, renal function, and contrast agent characteristics, making it impossible to achieve true individualized parameter optimization and increasing the risk of adverse reactions such as contrast-induced nephropathy.

[0005] In terms of vascular analysis, existing technologies oversimplify the assessment of vascular anatomy and lack precise analysis of microstructures such as vessel diameter distribution, tortuosity, and branching characteristics. In particular, the quantification of hemodynamic characteristics is insufficient. This makes it impossible to accurately assess the actual tolerance of different vascular regions, increasing the risk of vascular injury.

[0006] Physiological status monitoring systems are mostly static assessments and lack dynamic risk assessment mechanisms. They are unable to capture the fluctuation patterns of physiological parameters such as respiration, heart rate, and blood oxygen in real time, and their impact on the angiography process, making it difficult to identify potential risk states in a timely manner.

[0007] In terms of safety assessment, existing technologies mostly use single dimensions or simple threshold judgments, lacking the ability to integrate multi-source data and conduct comprehensive risk assessments. They are unable to establish correlation models between injection pressure, physiological status, and historical reactions, resulting in one-sided safety assessment results and difficulty in accurately predicting individualized risks.

[0008] Injection control strategies generally rely on preset schemes or simple feedback adjustments, lacking intelligent parameter optimization algorithms and dynamic control mechanisms. This makes it impossible to precisely control the flow of blood based on the patient's real-time status and vascular characteristics, compromising angiographic effectiveness and failing to maximize patient safety. In light of this, the present invention proposes a high-pressure injection system for angiography to address these issues. Summary of the Invention

[0009] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solutions:

[0010] An angiography high-pressure injection system, comprising:

[0011] A data acquisition module is used to collect physiological data and medical records of target patients to obtain corresponding patient physiological data information; the patient physiological data information includes standardized patient information, physiological monitoring data, renal function assessment data and historical angiography examination records;

[0012] The parameter calculation module obtains the corresponding individualized contrast agent injection parameter coefficient based on standardized patient information and renal function assessment data and combined with the contrast agent type;

[0013] The vascular analysis module is used to obtain the vascular anatomical structure information corresponding to the target patient's vascular region, perform pressure assessment based on hemodynamic characteristics, and obtain the vascular pressure capacity corresponding to each vascular sub-region within the corresponding vascular region; and combine the patient's physiological data information to obtain the injection pressure adjustment factor corresponding to each vascular sub-region;

[0014] The physiological assessment module assesses the physiological status of the target patient's vascular area based on the patient's physiological data information, obtains the corresponding physiological status influencing factors, and combines the pre-established physiological status safety threshold to perform risk marking and obtain the corresponding physiological status risk marking values ​​at different times;

[0015] A safety construction module is used to fuse multi-source data based on the obtained injection pressure adjustment factor, physiological state risk marker value and historical angiography examination records, and to construct the angiography safety degree corresponding to the corresponding vascular sub-region based on the obtained data;

[0016] An injection control module is used to obtain the optimal parameter path for contrast agent injection corresponding to the target patient based on the obtained contrast safety and in combination with a deep reinforcement learning algorithm, and to control the high-pressure injection device to perform the injection operation based on the optimal parameter path;

[0017] Furthermore, the process of obtaining the patient's physiological data information includes:

[0018] Setting up a data collection unit, crawling data based on the data collection unit, and obtaining corresponding basic archival information and historical angiography examination records;

[0019] Digitally process and standardize the storage of collected basic archival information to obtain corresponding standardized patient information;

[0020] A data monitoring unit is provided, and physiological parameters of the patient are collected in real time based on the data monitoring unit; corresponding physiological parameter data is obtained; and data filtering and noise elimination processing are performed on the collected physiological parameters to obtain smooth physiological monitoring data;

[0021] The data monitoring unit is also used to collect data on the patient's renal function related indicators to obtain corresponding renal function assessment data;

[0022] The obtained standardized patient information, physiological monitoring data, renal function assessment data and historical angiography examination records are summarized to obtain the corresponding patient physiological data information.

[0023] Furthermore, the process of obtaining the individualized contrast agent injection parameter coefficients includes:

[0024] Obtaining a physical health index corresponding to the target patient based on the standardized patient information; and obtaining a weight range corresponding to the target patient based on the physical health index;

[0025] Perform benchmark dose mapping based on the target patient's weight range to obtain a benchmark dose value corresponding to the target patient's weight; and perform unit weight injection volume adjustment and upper limit control on the obtained benchmark dose value to obtain a corresponding weight adjustment coefficient;

[0026] Obtain age grouping results for the corresponding patients based on standardized patient information; and perform physiological function attenuation estimation and compensation factor calculation based on the age grouping results to obtain the corresponding age correction coefficient;

[0027] The collected renal function assessment data are classified into renal function grades and risk assessments to obtain the corresponding renal function status index;

[0028] Based on the obtained renal function status index, a corresponding renal load assessment model is constructed, and the renal function risk level of the corresponding patient is divided into risk levels based on the renal function assessment data to obtain the corresponding renal function risk level;

[0029] Predict contrast agent clearance and set safety thresholds based on the patient's corresponding renal function risk level; obtain the corresponding renal function protection coefficient;

[0030] Obtain contrast agent type information; perform fluid dynamics analysis and injection characteristic simulation based on the contrast agent type information to obtain corresponding contrast agent characteristic correction factors;

[0031] The weight adjustment coefficient, age correction coefficient, renal function protection coefficient and contrast agent characteristic correction factor are used as input variables; and weights are assigned to each input variable. By applying a weighted fusion algorithm, each input variable is integrated to obtain the corresponding individualized contrast agent injection parameter coefficient.

[0032] Furthermore, the process of obtaining the vascular anatomical structure information corresponding to the target patient's vascular region includes:

[0033] Performing initial impact acquisition of the patient's vascular region based on a pre-selected angiography device to obtain corresponding original vascular images;

[0034] Performing digital enhancement and noise suppression on the original vascular images to obtain corresponding optimized vascular images;

[0035] Performing edge detection and contour extraction on the obtained optimized blood vessel image to obtain corresponding blood vessel contour data;

[0036] Obtaining the corresponding vascular diameter of each blood vessel in the target vascular region based on the extracted vascular contour data; and generating a corresponding vascular diameter distribution map based thereon;

[0037] Performing three-dimensional reconstruction based on the obtained blood vessel diameter distribution map and the blood vessel diameter distribution map to obtain a spatial structure model corresponding to the corresponding blood vessel region;

[0038] Based on the spatial structure model, the curvature is quantitatively calculated and the bending points are identified to obtain the corresponding vascular curvature index set;

[0039] By applying a topological analysis algorithm, branch points within the corresponding spatial structure model are identified and the angles between the branch vessels and the main vessels at the corresponding branch points are measured. At the same time, the three-dimensional coordinates of the corresponding branch points are recorded. Based on the three-dimensional coordinates and angles of the branch points, the corresponding vascular branch topology is constructed.

[0040] Based on the obtained vascular branch topology and the supply tissue range of the branch vessels corresponding to the branch points, the importance of the branch vessels is evaluated, and the blood flow ratio in the corresponding branch vessels is estimated by applying a pre-built fluid dynamics model. The obtained blood flow ratio and importance evaluation results are summarized to obtain the corresponding branch characteristic data;

[0041] Identifying and extracting anatomical feature points corresponding to the target vascular region based on the obtained vascular diameter distribution map, vascular tortuosity index set, and branch characteristic data; and dividing the corresponding vascular region into a plurality of continuous vascular sub-regions based on the anatomical feature points;

[0042] The anatomical index structure of the corresponding vascular sub-region is established to form complete vascular anatomical structure information.

[0043] Furthermore, obtaining the vascular pressure bearing capacity corresponding to each vascular sub-region includes:

[0044] The blood velocity and blood flow volume corresponding to each vascular sub-region are measured in combination with the obtained anatomical structure information; a three-dimensional reconstruction is performed based on the measurement results to obtain a corresponding blood velocity vector field; and the Reynolds number corresponding to each sub-region is obtained based on the obtained blood velocity vector field; based on the obtained blood velocity vector field, a corresponding hemodynamic characteristic index set is obtained;

[0045] Identify the corresponding optimized vascular images, obtain the vascular wall thickness in each vascular sub-region, and construct a corresponding wall thickness distribution map based on it;

[0046] Statistical analysis is performed on the obtained wall thickness distribution map to obtain corresponding statistical feature information; anomaly detection algorithms are used to identify abnormal points in the wall thickness; corresponding vascular wall structural features are obtained; internal pressure loads are applied based on the vascular wall structural features to obtain corresponding stress-strain curves; and wall tension is calculated based on the stress-strain curves to obtain corresponding elastic modulus distributions;

[0047] The maximum tolerable pressure is estimated based on the obtained elastic modulus distribution to obtain the corresponding theoretical upper limit of pressure;

[0048] Obtain target patient's age, vascular disease history and other vascular health data based on standardized patient information; perform vascular wall strength correction and safety factor setting based on vascular health data; and obtain corresponding vascular status adjustment factors;

[0049] The vascular pressure bearing capacity of each vascular sub-region is obtained based on hemodynamic characteristic indicators, vascular wall structural characteristics, theoretical pressure upper limit and vascular state adjustment factor.

[0050] Furthermore, the process of obtaining the injection pressure adjustment factor includes:

[0051] Based on the collected physiological monitoring data, the heart rate change curve corresponding to the target patient is obtained, and the heart rate variability analysis and periodic feature extraction are performed to obtain the corresponding heart rate fluctuation pattern;

[0052] Based on the heart rate fluctuation pattern, the cardiac systolic and diastolic cycles are divided and the time window is identified to obtain the optimal injection timing index;

[0053] Extracting statistical features from the systolic and diastolic blood pressure data of the corresponding patient based on the physiological monitoring data to obtain corresponding blood pressure state features;

[0054] Obtain the vascular pressure bearing capacity of each vascular sub-region and construct a pressure safety threshold curve based on it; set the injection pressure upper limit and risk level based on the pressure safety threshold curve to obtain a sub-region pressure limit table;

[0055] Construct a three-dimensional relationship model of heart rate, blood pressure, and pressure bearing capacity, and conduct dynamic simulation analysis and sensitivity testing based on it to obtain a parameter sensitivity matrix;

[0056] Optimize the pressure regulation strategy and design a smooth transition based on the parameter sensitivity matrix to obtain the corresponding pressure regulation benchmark value;

[0057] The corresponding injection pressure adjustment factor is generated according to the optimal injection timing index, the internal pressure fluctuation range, the sub-area pressure limit table and the pressure adjustment reference value.

[0058] Furthermore, the process of obtaining the physiological state influencing factors includes:

[0059] Extract features from the collected physiological monitoring data to obtain the corresponding breathing pattern features of the corresponding patients;

[0060] Respiratory cycle identification and intrathoracic pressure change estimation are performed based on respiratory pattern characteristics to obtain the respiratory influence coefficient;

[0061] Based on the collected physiological monitoring data, the target patient's ECG waveform is obtained, and waveform features and abnormal pattern recognition are performed to obtain the corresponding heart rhythm state assessment results. Based on the heart rhythm state assessment results, cardiac output changes are predicted and hemodynamic impact analysis is performed to obtain the heart function impact coefficient;

[0062] Perform trend analysis and threshold monitoring on changes in blood oxygen saturation within physiological monitoring data to obtain corresponding oxygenation status data; perform tissue perfusion assessment and hypoxia risk prediction based on oxygenation status data to obtain the oxygenation impact coefficient;

[0063] Based on the obtained vascular anatomical structure information, a regional feature description corresponding to the target vascular sub-region is obtained; based on the regional feature description, risk weight allocation and key point identification are performed to obtain a regional sensitivity coefficient;

[0064] The physiological state influence factors at each moment are obtained by integrating the respiratory influence coefficient, cardiac function influence coefficient, oxygenation influence coefficient and regional sensitivity coefficient.

[0065] Furthermore, the process of obtaining the physiological state risk marker value includes:

[0066] Construct safety reference standards; make individual adjustments and special case corrections to the obtained safety reference standards to obtain the corresponding personalized safety threshold set;

[0067] Obtain the corresponding physiological state influencing factors at each moment, and perform component decomposition and parameter mapping based on them to obtain a multidimensional physiological index vector;

[0068] Compare the multidimensional physiological indicator vectors with the personalized safety threshold set one by one and calculate the deviation to obtain a parameter deviation list;

[0069] Quantify the risk level and assign weights based on the parameter deviation list to obtain a weighted risk score;

[0070] Risk classification and criticality identification are performed based on weighted risk scores to obtain risk classification results;

[0071] Based on the time series of the influencing factors of the current physiological state and the corresponding physiological monitoring parameters, time series risk prediction and trend analysis are performed to obtain a risk evolution trend chart;

[0072] Based on the risk evolution trend chart, the warning level is determined and the risk cumulative effect is evaluated to obtain a dynamic risk index;

[0073] The risk grading results and dynamic risk index are integrated, and the comprehensive risk assessment results are generated through fuzzy logic reasoning algorithm, which are then clarified to obtain the corresponding physiological status risk marker value.

[0074] Furthermore, the process of obtaining the angiographic safety degree includes:

[0075] Obtain the injection pressure regulation factor corresponding to each vascular sub-region, and perform spatial distribution analysis and key area identification based on it to obtain a pressure-sensitive area map;

[0076] Based on the pressure sensitive area map, risk level assessment and safety boundary delineation are performed to obtain a pressure safety score;

[0077] Organize the physiological state risk marker values ​​at each moment into a time series, and conduct time series pattern mining and fluctuation characteristic analysis to obtain the risk fluctuation characteristics;

[0078] Based on the description of risk fluctuation characteristics, stability assessment and abnormal point detection are performed to obtain physiological risk scores;

[0079] Feature extraction is performed on the obtained historical angiography examination records to obtain the corresponding historical angiography feature set; based on the historical angiography feature set, angiography tolerance analysis and sensitivity prediction are performed to obtain a historical risk score;

[0080] Establish a three-source data feature space, perform feature extraction and dimensionality reduction on the stress safety score, physiological risk score, and historical risk score to obtain a fused feature vector. Establishing the three-source data feature space refers to constructing a high-dimensional feature representation framework encompassing the three dimensions of stress, physiology, and history.

[0081] A multi-level safety assessment model is constructed based on the fusion feature vector to obtain preliminary safety assessment results;

[0082] The obtained fusion feature vector is input into the constructed multi-level security assessment model to generate a multi-dimensional preliminary security assessment result;

[0083] The preliminary safety assessment results are calibrated and optimized to obtain the angiographic safety of the template patient in the corresponding vascular sub-region.

[0084] Furthermore, the process of obtaining the optimal parameter path for contrast agent injection includes:

[0085] Obtain angiography safety data and pre-set imaging quality requirement parameters for each vascular sub-region, and construct corresponding optimization target constraint conditions based on them;

[0086] Based on the optimization target constraints, the mapping relationship between injection parameters, safety and imaging quality is established, and the corresponding parameter-effect response surface is constructed based on it;

[0087] Constructing a corresponding parameter state space based on contrast injection parameters corresponding to the corresponding contrast injection process;

[0088] Define the action set and state transition function in the parameter space to obtain the injection control Markov decision process model;

[0089] Based on the parameter state space, action set and state transition function, a complete model of Markov decision process represented by four tuples is constructed;

[0090] Design a deep reinforcement learning reward function and train a deep Q network or policy gradient network based on it to obtain the corresponding injection strategy model; use online learning methods to adjust and optimize the injection strategy model in real time to obtain an adaptive injection control strategy;

[0091] Based on the adaptive injection control strategy, path planning and trajectory optimization are performed in the parameter state space to obtain the injection parameter time series;

[0092] The injection parameter time series is subjected to safety verification and robustness testing to obtain the optimal parameter path for the corresponding contrast agent injection; based on this, the corresponding injection control instructions are generated and transmitted to the high-pressure injection equipment for execution.

[0093] The technical effects and advantages of the angiography high-pressure injection system of the present invention are as follows:

[0094] 1. Combining multidimensional analysis of vascular diameter distribution maps, vascular tortuosity index sets, and branch characteristic data, this approach obtains vascular anatomical structure information and hemodynamic characteristics, enabling accurate assessment of the capacity of each vascular subregion and effectively preventing the risk of vascular injury.

[0095] 2. Through real-time monitoring and integration of respiratory influence coefficients, cardiac function influence coefficients, oxygenation influence coefficients, and regional sensitivity coefficients, a dynamic physiological state assessment model was constructed, enabling real-time labeling and early warning of multi-dimensional physiological risks, enabling the system to promptly identify and respond to potential dangerous conditions;

[0096] 3. Based on multi-source data fusion technology using stress safety scores, physiological risk scores, and historical risk scores, a comprehensive angiography safety assessment system has been established. This system provides quantitative indicators that comprehensively reflect the safety status of individual patients and enables accurate risk prediction and hierarchical management.

[0097] 4. Through deep reinforcement learning algorithms and Markov decision process models, intelligent planning and dynamic adjustment of the optimal parameter path for contrast agent injection are achieved, maximizing patient safety while ensuring imaging quality, significantly improving the overall efficiency and quality of angiography examinations. BRIEF DESCRIPTION OF THE DRAWINGS

[0098] Figure 1 This is a schematic diagram of a high-pressure injection system for angiography according to the present invention. DETAILED DESCRIPTION

[0099] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0100] Example 1

[0101] See also Figure 1 As shown, the angiography high-pressure injection system described in this embodiment includes:

[0102] A data acquisition module is used to collect physiological data and medical records of target patients to obtain corresponding patient physiological data information; the patient physiological data information includes standardized patient information, physiological monitoring data, renal function assessment data and historical angiography examination records;

[0103] The parameter calculation module obtains the corresponding individualized contrast agent injection parameter coefficient based on standardized patient information and renal function assessment data and combined with the contrast agent type;

[0104] The vascular analysis module is used to obtain the vascular anatomical structure information corresponding to the target patient's vascular region, perform pressure assessment based on hemodynamic characteristics, and obtain the vascular pressure capacity corresponding to each vascular sub-region within the corresponding vascular region; and combine the patient's physiological data information to obtain the injection pressure adjustment factor corresponding to each vascular sub-region;

[0105] The physiological assessment module assesses the physiological status of the target patient's vascular area based on the patient's physiological data information, obtains the corresponding physiological status influencing factors, and combines the pre-established physiological status safety threshold to perform risk marking and obtain the corresponding physiological status risk marking values ​​at different times;

[0106] A safety construction module is used to fuse multi-source data based on the obtained injection pressure adjustment factor, physiological state risk marker value and historical angiography examination records, and to construct the angiography safety degree corresponding to the corresponding vascular sub-region based on the obtained data;

[0107] An injection control module is used to obtain the optimal parameter path for contrast agent injection corresponding to the target patient based on the obtained contrast safety and in combination with a deep reinforcement learning algorithm, and to control the high-pressure injection device to perform the injection operation based on the optimal parameter path;

[0108] It should be further explained that, in the specific implementation process, the process of obtaining the patient's physiological data information includes:

[0109] Setting up a data collection unit, wherein the data collection unit is connected to an information resource library within a corresponding medical institution through an API interface; the information resource library includes but is not limited to an electronic medical record information library, a pharmaceutical knowledge library, etc.;

[0110] The data acquisition unit crawls the corresponding information resource library based on the API interface to obtain corresponding basic file information and historical angiography examination records; the basic file information includes basic information such as the patient's age, height, weight and previous medical history; the historical angiography examination records include data such as the patient's historical contrast agent usage and reaction status;

[0111] The collected basic archival information is digitized and stored in a standardized manner to obtain corresponding standardized patient information. The digitization includes processes such as digital format conversion and numerical unit unification. The standardized storage is to structure the digitized basic archival information using a unified coding standard to ensure consistency in field naming, units, and value ranges.

[0112] At the same time, a data monitoring unit is set up, and the patient's physiological parameters are collected in real time based on the data monitoring unit to obtain corresponding physiological parameter data; the physiological parameter data include time series data corresponding to physiological parameters such as heart rate, blood pressure, respiratory rate and blood oxygen saturation; and the collected physiological parameters are subjected to data filtering and noise elimination processing to obtain smooth physiological monitoring data; the data filtering refers to the removal of signal interference and noise generated during the physiological parameter collection process based on a pre-selected low-pass filter; the noise removal refers to the identification and deletion of abnormal data points caused by factors such as equipment failure, patient movement or external interference through a median filtering algorithm.

[0113] At the same time, the data monitoring unit is also used to collect data on the patient's renal function-related indicators to obtain corresponding renal function assessment data, including creatinine, urea nitrogen and estimated glomerular filtration rate;

[0114] The obtained standardized patient information, physiological monitoring data, renal function assessment data and historical angiography examination records are summarized to obtain the corresponding patient physiological data information.

[0115] It should be further explained that, in the specific implementation process, the process of obtaining the individualized contrast agent injection parameter coefficient includes:

[0116] Obtaining a body mass index (BMI) corresponding to the target patient based on the standardized patient information, wherein the body mass index represents the ratio of the patient's weight to the square of his or her height; and obtaining a weight range corresponding to the target patient based on the body mass index (BMI), wherein the weight range includes a low weight range (BMI < 18.5), a standard weight range (18.5 ≤ BMI ≤ 24), and a high weight range (BMI > 24);

[0117] A benchmark dose mapping is performed based on the target patient's weight range; the benchmark dose mapping refers to establishing a basic dose mapping table based on clinical medical guidelines and in combination with different contrast agent types and iodine concentrations; and based on it, a benchmark dose value corresponding to the target patient's weight is obtained; and the obtained benchmark dose value is adjusted for injection volume per unit weight and an upper limit control is performed on the obtained benchmark dose value to obtain a corresponding weight adjustment coefficient; wherein, the injection volume per unit weight adjustment refers to adjusting the obtained benchmark dose value based on the examination type and vascular area required by the patient; for example: if the examination type is CT pulmonary angiography, and the corresponding vascular area is relatively thin; the obtained benchmark dose value needs to be adaptively increased; the upper limit control value is based on the maximum benchmark dose value allowed by clinical medical guidelines to prevent overinjection; the weight adjustment coefficient refers to the ratio between the benchmark dose values ​​before and after adjustment;

[0118] Based on the standardized patient information, the age grouping results corresponding to the corresponding patients are obtained, and the age groups include children, youth, middle-aged and elderly people; and based on the age grouping results, physiological function attenuation is estimated and compensation factors are calculated to obtain the corresponding age correction coefficient; the functional attenuation estimation refers to obtaining the physiological characteristics corresponding to the target patient based on the patient's physiological data information, and mathematically modeling is performed based thereon to obtain the corresponding physiological function attenuation model, and based on the model, the functional attenuation index corresponding to the corresponding patient is obtained; the physiological function attenuation model is used to quantify the degree of influence of aging on the patient's renal clearance function, cardiac function and vascular elasticity changes; the compensation factor calculation is based on the functional attenuation index to compensate for the corresponding patient's metabolism and excretion process of the contrast agent; for example: for elderly patients, due to the decline in their physiological function due to aging, the dosage and injection rate of the contrast agent need to be adjusted to adapt to the patient's reduced metabolic and excretion capacity; and then, the corresponding age correction coefficient is obtained based on the physiological function attenuation estimation and compensation factor calculation results;

[0119] The collected renal function assessment data are subjected to renal function grade classification and risk assessment to obtain a corresponding renal function status index; the renal function grade classification refers to the classification of the patient's renal function based on the KDIGO standard and the assignment of grade scores; for example, there are five grades: normal, mild decline, moderate decline, moderate and severe decline, and renal failure; the KDIGO standard represents the clinical practice guidelines developed by the Kidney Disease Improving Global Outcomes Organization; the risk assessment refers to the contrast agent nephropathy risk score considering the patient's age, previous medical history, and renal function grade classification results to form a corresponding renal function status index; and the lower the renal function status index, the better the renal function status of the corresponding patient and the lower the risk of contrast agent use;

[0120] Based on the obtained renal function status index, a corresponding renal load assessment model is constructed, and the renal function risk level of the corresponding patient is divided into low risk, medium risk and high risk levels by using the renal load assessment model and combining the renal function assessment data;

[0121] Furthermore, the contrast agent clearance rate is predicted and the safety threshold is set according to the patient's corresponding renal function risk level; the contrast agent clearance prediction formula is based on pharmacokinetic theory, and the clearance rate and half-life of the contrast agent in the target patient are predicted by comprehensively considering the influence of physiological function attenuation; the safety threshold setting refers to adjusting the baseline injection dose upper limit and injection rate limit of the contrast agent based on the contrast agent clearance prediction results and clinical medicine guidelines; further, the corresponding renal function protection coefficient is obtained based on the contrast agent clearance prediction results and the safety threshold setting results; the process of obtaining the renal function protection coefficient includes: comparing the adjusted baseline injection dose upper limit and injection rate limit of the contrast agent with the maximum allowed baseline injection dose and injection rate, obtaining the corresponding ratio results and summing them to obtain the corresponding safety threshold compensation coefficient; similarly, comparing the predicted contrast agent clearance rate with the pre-set standard contrast agent clearance rate to obtain the corresponding ratio result, and using it as the clearance compensation factor; further, adding the clearance compensation factor, the safety threshold compensation coefficient and the pre-set baseline renal function protection coefficient to obtain the corresponding renal function protection coefficient;

[0122] Information retrieval is performed on the corresponding pharmaceutical knowledge base to obtain corresponding contrast agent type information, which includes contrast agent physical and chemical characteristic parameters such as iodine concentration, viscosity, and osmotic pressure data. Based on the contrast agent type information, fluid dynamics analysis and injection characteristic simulation are performed to obtain the corresponding contrast agent characteristic correction factor, which reflects the impact of the contrast agent's physical and chemical characteristics on the injection parameters. Fluid dynamics analysis refers to quantifying the flow resistance, viscosity change, and pressure flow rate changes of the contrast agent under different iodine concentration and temperature conditions through Poiseuille's law. Injection characteristic simulation is to use computer simulation to predict the flow behavior of the contrast agent in the injection device and blood vessels. The flow behavior includes laminar / turbulent flow conversion and pressure wave propagation to evaluate the performance differences of different types of contrast agents during the injection process.

[0123] The weight adjustment coefficient, age correction coefficient, renal function protection coefficient and contrast agent characteristic correction factor are used as input variables; and weights are assigned to each input variable; each input variable is integrated by applying a weighted fusion algorithm to obtain a corresponding individualized contrast agent injection parameter coefficient; the individualized contrast agent injection parameter coefficient can be used to adjust the standard injection plan to achieve individualized customization of contrast agent dosage, injection rate and pressure, thereby improving the safety and effectiveness of angiographic examinations.

[0124] It should be further explained that, in a specific implementation process, the process of obtaining the vascular anatomical structure information corresponding to the target patient's vascular region includes:

[0125] Performing initial image acquisition of the patient's vascular region based on a pre-selected angiography device to obtain a corresponding original vascular image; the angiography device is one of a digital subtraction angiography device, a CT angiography device, or a magnetic resonance angiography device; the original vascular image includes vascular distribution and morphological information of the target region;

[0126] Then, the original vascular image is digitally enhanced and noise suppressed to obtain a corresponding optimized vascular image. The digital enhancement refers to contrast enhancement, brightness adjustment, and edge sharpening of the acquired original vascular image to improve the visibility of the vascular structure. The noise suppression eliminates noise in the original vascular image after data enhancement through an adaptive filtering algorithm, which can be a Gaussian filter or a median filter, while retaining the detailed information of the vascular structure.

[0127] Performing edge detection and contour extraction on the obtained optimized vascular image; the edge detection refers to identifying edge information of the vascular region by using the Canny edge detection algorithm; the contour extraction refers to extracting the vascular contour of the corresponding vascular region based on the edge detection result and using a pre-built contour model to obtain the corresponding vascular contour data;

[0128] Furthermore, the vascular diameters of the respective blood vessels within the target vascular region are obtained based on the extracted vascular contour data, and the rate of change of the vascular diameters along the length of the corresponding blood vessels is calculated based on the obtained data to identify stenosis and dilation regions within the corresponding vascular region; and a corresponding vascular diameter distribution map is generated based on the obtained data.

[0129] Based on the obtained blood vessel diameter distribution map and the blood vessel diameter distribution map, a three-dimensional reconstruction is performed to obtain a spatial structure model corresponding to the corresponding blood vessel area; the three-dimensional reconstruction adopts multi-view reconstruction technology to construct a three-dimensional spatial grid model of the blood vessel;

[0130] Performing curvature quantification and bend point identification based on the spatial structure model; obtaining a corresponding vascular tortuosity index set; the curvature quantification is based on the local curvature and torsion corresponding to each node in the spatial structure model, quantifying the degree of curvature corresponding to the corresponding node; the bend point identification identifies nodes whose curvature exceeds a preset curvature threshold based on the curvature quantification result and marks them as high curvature points; and based on this, obtaining a corresponding vascular tortuosity index set; the vascular tortuosity index set includes the location, local curvature, and torsion of each high curvature point;

[0131] Furthermore, a topological analysis algorithm is applied to identify branch points within the corresponding spatial structure model, and the angles between the branch vessels and the main vessels at the corresponding branch points are measured. The three-dimensional coordinates of the corresponding branch points are recorded, and the corresponding vascular branch topology is constructed based on the three-dimensional coordinates and angles of the branch points.

[0132] Based on the obtained vascular branch topology and the supply tissue range of the branch vessels corresponding to the branch points, the importance of the branch vessels is evaluated, and the blood flow ratio in the corresponding branch vessels is estimated by applying a pre-built fluid dynamics model. The obtained blood flow ratio and importance evaluation results are summarized to obtain the corresponding branch characteristic data;

[0133] Then, based on the obtained vascular diameter distribution map, vascular tortuosity index set, and branching characteristic data, anatomical feature points corresponding to the target vascular region are identified and extracted. The anatomical feature points include stenosis points, dilation points, high curvature points, and branching points. The corresponding vascular region is then divided into a number of continuous vascular sub-regions based on the anatomical feature points. Each vascular sub-region has relatively consistent anatomical structure and blood flow characteristics.

[0134] An anatomical index structure is established for the corresponding vascular sub-region, which includes information such as the location, length, average diameter, curvature, branching, and blood supply range of the vascular sub-region, thereby forming complete vascular anatomical structure information.

[0135] It should be further explained that, in the specific implementation process, the acquisition of the vascular pressure bearing capacity corresponding to each vascular sub-region includes:

[0136] Doppler ultrasound technology is used, combined with the obtained anatomical structure information, to measure the blood flow velocity and blood flow volume corresponding to each vascular sub-region; three-dimensional reconstruction is performed based on the measurement results to obtain a corresponding blood flow velocity vector field; the blood flow velocity vector field is used to describe the velocity magnitude and direction of each branch point in the corresponding vascular sub-region; and the Reynolds number corresponding to each sub-region is obtained based on the obtained blood flow velocity vector field; the Reynolds number is used to measure the distribution characteristics of blood flow; the distribution characteristics include laminar flow, transitional flow, and turbulent flow; and based on the distribution characteristics, hemodynamic parameters such as wall shear stress and vorticity at locations where high pressure and fluid disturbance may occur are identified and extracted to obtain a corresponding hemodynamic characteristic index set;

[0137] By using high-resolution vascular imaging technology to identify the corresponding optimized vascular images, the vascular wall thickness in each vascular sub-region is obtained; and based on the obtained information, a corresponding wall thickness distribution map is constructed, which is used to reflect the thickness changes of the vascular wall at different locations;

[0138] Statistical analysis is performed on the obtained wall thickness distribution map to obtain corresponding statistical feature information; the statistical feature information includes statistical features such as average wall thickness and wall thickness standard deviation; and anomaly detection algorithms are used to identify abnormal wall thickness points, including areas of partial thinning or thickening (e.g., plaques or tumor-like expansions), to obtain corresponding vascular wall structural features;

[0139] The vascular wall is simulated as a nonlinear, anisotropic elastic material based on a finite element analysis algorithm, and an internal pressure load is applied to simulate the stress-strain relationship of the blood vessel under different internal pressure loads, thereby obtaining corresponding stress-strain curves. Based on this, wall tension is calculated to obtain the corresponding elastic modulus distribution. The wall tension calculation refers to the quantification of the tension level of the blood vessel wall based on Laplace's law. The elastic modulus distribution includes the tangential elastic modulus and the axial elastic modulus.

[0140] Furthermore, the maximum tolerable pressure is estimated based on the obtained elastic modulus distribution to obtain the corresponding theoretical upper pressure limit. The maximum tolerable pressure estimation is achieved by simulating the degree of expansion of blood vessels under different pressures and predicting the theoretical upper pressure limit of each vascular sub-region according to clinical safety guidelines.

[0141] Based on standardized patient information, vascular health data such as the target patient's age and history of vascular disease are obtained; based on the vascular health data, vascular wall strength correction and safety factor setting are performed to obtain a corresponding vascular status adjustment factor; the vascular wall correction refers to the assessment of the patient's vascular elasticity, thickness and structural integrity based on the vascular health data, and the adjustment of the theoretical upper limit of the vascular wall pressure based on the assessment results; the safety factor setting reserves a safety margin based on the theoretical maximum pressure value to prevent accidental damage; and then, through the process of vascular wall strength correction and safety factor setting, the corresponding vascular status adjustment factor is obtained;

[0142] Furthermore, by integrating hemodynamic characteristic indicators, vascular wall structural characteristics, theoretical pressure upper limit and vascular state adjustment factors, the safe working pressure range of each vascular sub-region is calculated through a multi-objective optimization algorithm to obtain the vascular pressure capacity of each vascular sub-region; the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization algorithm.

[0143] It should be further explained that, in the specific implementation process, the process of obtaining the injection pressure adjustment factor includes:

[0144] Based on the collected physiological monitoring data, the heart rate change curve corresponding to the target patient is obtained, and heart rate variability analysis and periodic feature extraction are performed on the heart rate variability to obtain the corresponding heart rate fluctuation pattern, which includes normal sinus rhythm, respiratory heart rate variability, and non-periodic fluctuations. The heart rate fluctuation model is used to reflect the stability and regularity of the cardiac contraction rhythm. The heart rate variability analysis refers to obtaining the stability and fluctuation characteristics corresponding to the corresponding heart rate change curve by combining time domain analysis and frequency domain analysis. The periodic feature extraction uses Fourier transform to identify the periodic pattern of heart rate variation.

[0145] Based on the heart rate fluctuation pattern, cardiac systolic and diastolic cycle division and time window identification are performed to obtain the optimal injection timing indicator. Cardiac systolic and diastolic cycle division refers to dividing the cardiac cycle into systolic and diastolic phases through R wave detection. Time window identification is used to identify the optimal injection parameters corresponding to different cardiac cycle phases and form the optimal injection time indicator based on them.

[0146] Based on the physiological monitoring data, statistical feature extraction is performed on the systolic and diastolic blood pressure data of the corresponding patient to obtain corresponding blood pressure status characteristics; the blood pressure status characteristics include mean arterial pressure and pulse pressure difference; and based on the obtained blood pressure status characteristics, intravascular pressure estimation and dynamic change prediction are performed to obtain the corresponding internal pressure fluctuation range; the intravascular pressure estimation refers to estimating the actual internal pressure level at different branch points in the blood vessel based on the patient's systolic and diastolic blood pressure data, combined with the anatomical structure. The corresponding estimation process needs to consider factors such as the conduction characteristics of the blood pressure at the branch point, changes in blood vessel diameter and blood flow resistance; and the non-invasive measurement results are converted into pressure estimates at the target branch point; the dynamic change prediction predicts the pressure fluctuation range that may occur during the contrast agent injection process by analyzing the patient's blood pressure fluctuation pattern over time, combined with physiological factors such as respiratory cycle and heart rate variability;

[0147] Obtain the vascular pressure bearing capacity of each vascular sub-region and construct a pressure safety threshold curve based on it; the pressure safety threshold curve defines the safety factor of each vascular sub-region under different pressures;

[0148] Based on the pressure safety threshold curve, the injection pressure upper limit is set and the risk level is divided to obtain a sub-region pressure limit table. The injection pressure upper limit setting is used to ensure that the injection pressure during contrast agent injection does not exceed the pressure bearing capacity of the blood vessel. The risk level division is equivalent to dividing the injection pressure during the injection process into pressure range intervals of different risk levels and classifying them for management, thus forming a corresponding regional pressure limit table.

[0149] A three-dimensional relationship model of heart rate, blood pressure and pressure bearing capacity is constructed, and dynamic simulation analysis and sensitivity testing are performed to obtain a parameter sensitivity matrix. Dynamic simulation analysis refers to the real-time changes of blood vessels during contrast agent injection based on computer simulation. The simulation process takes into account factors such as changes in the cardiac cycle, vascular elastic response, and contrast agent flow characteristics, and generates time-varying curves of parameters such as vascular pressure, flow velocity, and wall stress. Sensitivity testing evaluates the impact of changes in each parameter on injection safety based on the change curve. Dynamic simulation analysis and sensitivity testing can quantify the degree of change in vascular pressure bearing capacity under different physiological states, and form a corresponding parameter sensitivity matrix based on it. For example, by changing the value of a physiological parameter alone and observing the magnitude of the change in vascular pressure bearing capacity, the sensitivity of the corresponding physiological parameter can be quantified.

[0150] The pressure regulation strategy is optimized and the smooth transition design is performed according to the parameter sensitivity matrix to obtain the corresponding pressure regulation reference value; the pressure regulation strategy optimization refers to constructing a multidimensional optimization model based on the parameter sensitivity matrix, and iteratively calculating the optimal pressure control strategy for each vascular sub-region by comprehensively considering safety and angiography effect; the smooth transition design refers to constructing a continuous and differentiable pressure change curve using spline interpolation and Lagrange polynomials, and the pressure change curve achieves shockless transition at the junction of vascular sub-regions, physiological state change points and injection stage transitions, and ensures the smoothness of pressure change through pressure change rate limitation and elastic buffer mechanism; by combining the optimization strategy with the smooth curve, a pressure regulation reference value including a reference pressure value, an allowable deviation range, an upper limit of the change rate and an emergency adjustment coefficient is formed;

[0151] According to the optimal injection timing index, internal pressure fluctuation range, sub-region pressure limit table and pressure adjustment baseline value, the contrast agent injection parameters are dynamically adjusted through the proportional-integral-differential (PID) controller to generate a comprehensive parameter matrix including the basic pressure value, dynamic adjustment coefficient, timing weight and safety limit factor, which is recorded as the injection pressure adjustment factor.

[0152] It should be further explained that, in the specific implementation process, the process of obtaining the physiological status influencing factors includes:

[0153] Extract features from the collected physiological monitoring data to obtain the corresponding breathing pattern features of the corresponding patients; the breathing pattern features describe the rhythm and intensity of the patient's breathing;

[0154] Respiratory cycle identification and intrathoracic pressure change estimation are performed based on respiratory pattern characteristics to obtain a respiratory influence coefficient; respiratory cycle identification refers to extracting the time domain and frequency domain features of respiratory data through signal processing algorithms, identifying key time nodes such as the start of inspiration, peak inspiration, start of exhalation, and end of exhalation, thereby constructing a complete respiratory cycle; intrathoracic pressure change estimation deduces the change of intrathoracic pressure over time through the respiratory depth, frequency, and rhythm parameters within the corresponding respiratory cycle, and based on this, obtains the degree of influence of the respiratory process on the change of intrathoracic pressure; the respiratory imaging coefficient refers to a comprehensive parameter that quantifies the degree of influence of the patient's respiratory state on the intravascular pressure and contrast agent distribution;

[0155] Based on the collected physiological monitoring data, an electrocardiogram (ECG) waveform of the target patient is obtained, and waveform features and abnormal pattern recognition are performed on the waveform to obtain a corresponding cardiac rhythm state assessment result; the waveform feature extraction refers to extracting waveform features by applying wavelet transform and morphological analysis methods, and the waveform features include the morphology, amplitude and time interval of the P wave, QRS complex wave and T wave; the abnormal pattern refers to identifying and extracting abnormal cardiac rhythm patterns such as premature beats, tachycardia and conduction block in the ECG waveform based on a machine learning algorithm; through waveform feature extraction and abnormal pattern recognition, a cardiac rhythm state assessment result including heart rate stability, cardiac rhythm type, conduction function and abnormal event statistics is generated;

[0156] Based on the results of cardiac rhythm state assessment, cardiac output change prediction and hemodynamic impact analysis are performed to obtain a cardiac function impact coefficient; the cardiac output change prediction refers to the estimation of cardiac output changes based on the heart rate and cardiac rhythm state assessment results; the cardiac output refers to the total amount of blood ejected by the left ventricle or right ventricle into the aorta or pulmonary artery per minute; the hemodynamic impact analysis is based on fluid mechanics methods, combined with cardiac output prediction results and spatial structure models, to simulate blood flow distribution, blood flow velocity and wall shear stress changes under different cardiac rhythm states, and based on this, the degree and duration of the impact of cardiac rhythm changes on the hemodynamic parameters of the target vascular area are evaluated; by integrating the results of cardiac output change prediction and hemodynamic impact analysis, and applying a weighted fusion algorithm, a cardiac function impact coefficient that can be used to quantify the impact of cardiac function on the angiography process is obtained;

[0157] Trend analysis and threshold monitoring are performed on changes in blood oxygen saturation within physiological monitoring data to obtain corresponding oxygenation status data, which can be used to reflect the blood's oxygen-carrying capacity and tissue oxygen supply. Trend analysis refers to identifying downward and fluctuating trends within the time series data corresponding to blood oxygen saturation. Threshold monitoring assesses the severity of hypoxemia in patients based on pre-set multi-level blood oxygen saturation thresholds and duration.

[0158] Based on oxygenation status data, tissue perfusion assessment and hypoxia risk prediction are performed to obtain an oxygenation impact coefficient. Tissue perfusion assessment refers to the use of blood oxygen saturation, pulse waveform, and heart rate variability data to construct a tissue oxygen supply and demand balance assessment system, and based on this, indicators such as oxygen delivery efficiency and perfusion heterogeneity are obtained, and based on this, the blood supply status and oxygenation level in the target vascular area are quantified. The hypoxia risk prediction is used to predict hypoxia risk based on the blood supply status and oxygenation level in the blood. By integrating the tissue perfusion assessment results and the hypoxia risk prediction results and applying a fuzzy logic reasoning algorithm, an oxygenation impact coefficient that can be used to reflect the degree of influence of the patient's oxygenation status on contrast agent injection is obtained.

[0159] Obtaining a regional feature description corresponding to the target vascular sub-region based on the obtained vascular anatomical structure information, wherein the regional feature description includes vascular location, functional importance, and adjacent organ relationship;

[0160] Based on the description of regional characteristics, risk weight allocation and key point identification are performed to obtain the regional sensitivity coefficient; risk weight allocation refers to the application of the hierarchical analysis process (AHP) and the Delphi expert scoring theory to construct a multi-level risk assessment matrix based on vascular location, physiological function importance and clinical risk statistical data. The multi-level risk assessment matrix assigns a normalized risk weight value to each vascular sub-region by comparing the clinical vulnerability, complication rate and functional damage consequences of different vascular sub-regions; the key point identification is based on vascular morphology analysis and hemodynamics, combined with computer vision algorithms to identify special structural points in the blood vessels. Special structural points include bifurcations, stenosis segments, tumor-like expansion areas and abnormal bending areas; a network analysis method based on graph theory is used to evaluate the topological importance and blood flow distribution influence of the corresponding special structural points in the vascular network, and the corresponding evaluation results are integrated with the corresponding risk weight matrix, and a nonlinear mapping function is used to quantify the regional sensitivity coefficient corresponding to each vascular sub-region; the regional sensitivity coefficient can be used to measure the sensitivity of each vascular sub-region to pressure changes;

[0161] Through multi-sensor data fusion technology, the respiratory influence coefficient, cardiac function influence coefficient, oxygenation influence coefficient and regional sensitivity coefficient are integrated to construct a real-time physiological state assessment model, and the physiological state influence factor at each moment is obtained; the multi-sensor data fusion technology refers to the application of Kalman filtering and Bayesian network framework to achieve time synchronization and feature layer fusion of heterogeneous physiological signals; the input parameters of the physiological state assessment model are the corresponding respiratory influence coefficient, cardiac function influence coefficient, oxygenation influence coefficient and regional sensitivity coefficient; the real-time physiological state assessment model realizes nonlinear mapping through a multi-layer perceptron, and at the same time introduces expert knowledge constraints to ensure that the model output conforms to physiological laws, thereby generating a physiological state influence factor that comprehensively reflects the degree of influence of the patient's current physiological state on the angiography process.

[0162] It should be further explained that, in the specific implementation process, the process of obtaining the physiological state risk marker value includes:

[0163] Constructing multi-parameter physiological safety thresholds based on clinical safety guidelines to obtain corresponding safety reference standards, including normal ranges and safety limits for physiological parameters such as respiratory rate, heart rate variability, blood oxygen saturation, and blood pressure fluctuation range;

[0164] The obtained safety reference standards are individually adjusted and modified for special circumstances to obtain a corresponding personalized safety threshold set; individualized adjustment refers to targeted adjustment of the constructed safety reference standards based on the patient's actual physiological parameters, and the adjustment process takes into account factors such as the patient's age stratification, body mass index differences, underlying disease conditions, and organ function reserves to generate a safety parameter matrix adapted to individual characteristics. The personalized parameter standards include values ​​such as heart rate variability range, blood pressure fluctuation limit, respiratory rate safety interval, and blood oxygen saturation lower limit; special situation corrections are based on the patient's special physiological state or pathological conditions to further refine the individualized parameters, including narrowing the heart rate tolerance range for patients with cardiovascular disease, lowering the blood pressure fluctuation limit for patients with renal insufficiency, setting special thresholds for oxygenation indicators for patients with respiratory diseases, and adjusting sensitivity parameters for patients with a history of contrast agent allergy, to form the final personalized safety threshold set;

[0165] Obtain the corresponding physiological state influencing factors at each moment, and perform component decomposition and parameter mapping based on them to obtain a multidimensional physiological indicator vector; component decomposition refers to separating the physiological state influencing factors into several independent components; parameter mapping is used to map the corresponding independent vectors into a standardized metric space to form a multidimensional physiological indicator vector;

[0166] The multidimensional physiological indicator vectors are compared one by one with the personalized safety threshold set and the deviation is calculated to obtain a parameter deviation list; one-to-one comparison means that the obtained multidimensional physiological indicator vectors are compared with the corresponding personalized safety threshold set respectively. The corresponding comparison process needs to consider the differentiated processing of parameter types. The differentiated processing of parameter types includes upper and lower limit judgments for single threshold parameters, range inclusion tests for interval parameters, and trend conformity assessments for time series parameters. The deviation calculation quantifies the degree of deviation of each physiological parameter based on the comparison results and generates a parameter deviation list containing deviation indicators such as deviation direction, deviation amplitude, deviation duration, and deviation rate;

[0167] Based on the parameter deviation list, the risk degree is quantified and weighted to obtain a weighted risk score. Risk degree quantification refers to converting each deviation indicator in the parameter deviation list into a standardized risk value. The quantification process needs to take into account factors such as the difference in the risk of deviation direction, the nonlinear risk mapping of the deviation amplitude, and the cumulative effect of the deviation duration. Based on different degrees of deviation, corresponding risk levels are set to generate the original risk score of each deviation indicator. Weight assignment is based on the clinical importance and physiological impact of each deviation indicator, and the importance score driven by the expert knowledge base is used to assign weights. Based on the weighted sum of the risk degree quantification and weight assignment results, a weighted risk score that can be used to reflect the patient's current overall risk status is obtained.

[0168] Risk classification and critical state identification are performed based on the weighted risk score to obtain the risk classification result; risk classification refers to setting a multi-level risk classification standard based on the weighted risk score, and dividing it into different level areas such as safety, attention, warning and danger based on clinical medical guidelines, and obtaining the level area to which the current weighted risk score belongs; critical state identification is based on the boundary condition analysis of the mutation point detection algorithm to accurately capture the critical state where the risk level is about to transition, and generate a risk classification result that includes comprehensive information such as the current risk level, critical state sign, risk change trend and warning time window;

[0169] Based on the time series corresponding to the current physiological status influencing factors and physiological monitoring parameters, time series risk prediction and trend analysis are performed to obtain a risk evolution trend chart. Time series risk prediction refers to predicting the risk status by applying the sliding time window technology to generate the risk change trend at multiple time points in the future. The trend analysis refers to evaluating the direction and speed of the corresponding risk change trend based on the time series analysis algorithm to generate the corresponding risk evolution trend chart.

[0170] Based on the risk evolution trend chart, early warning level determination and risk accumulation effect assessment are performed to obtain a dynamic risk index. Early warning level determination refers to a multi-level early warning analysis of the risk change trend in the risk evolution trend chart, and a hierarchical early warning algorithm is used to divide the predicted risk change trend into different warning levels, such as routine monitoring, close observation, early warning, and emergency intervention. An early warning decision table is generated, which includes the warning level, trigger conditions, and recommended intervention timing. Risk accumulation effect assessment quantifies risk accumulation based on the risk change trend, integrates the early warning decision table and risk accumulation quantification results, and generates a dynamic risk index that reflects the real-time risk status and change trend.

[0171] The risk grading results and dynamic risk index are integrated, and the comprehensive risk assessment results are generated through fuzzy logic reasoning algorithm, which are then clarified to obtain the corresponding physiological status risk marker value.

[0172] It should be further explained that, in the specific implementation process, the process of obtaining the angiographic safety degree includes:

[0173] The injection pressure regulation factor corresponding to each vascular sub-region is obtained, and spatial distribution analysis and key area identification are performed based on it to obtain a pressure-sensitive area map. Spatial distribution analysis refers to the spatial mapping and distribution feature extraction of the injection pressure regulation factor based on the vascular sub-region to form a corresponding spatial distribution pattern. Key area identification is based on the spatial distribution pattern and uses computer vision technology to accurately locate abnormal areas and risk hotspots, accurately marking local areas with significantly higher pressure sensitivity than surrounding areas and potential risk points, and generating a pressure-sensitive area map that includes the location, size, sensitivity level and adjacent relationship of sensitive areas.

[0174] Based on the pressure-sensitive area map, risk level assessment and safety boundary delineation are performed to obtain a pressure safety score. Risk level assessment refers to the quantitative analysis and risk classification of the pressure sensitivity of each area in the pressure-sensitive area map, and the acquisition of the corresponding regional risk index for each area. Safety boundary delineation is based on computer image processing algorithms, combined with the risk level assessment results to accurately define the safe operating range, generate a safe operating boundary with a buffer zone design, integrate the regional risk index and safe operating boundary, and use a multi-objective optimization algorithm to obtain a pressure safety score that can be used to reflect the pressure safety status.

[0175] The physiological state risk marker values ​​at each moment are organized into time series, and time series pattern mining and fluctuation characteristic analysis are performed to obtain risk fluctuation characteristics; time series pattern mining refers to in-depth analysis and pattern recognition of the time series data corresponding to the physiological state risk marker values ​​to obtain the time series pattern of the corresponding physiological state risk marker values. The time series pattern includes time series information such as periodic characteristics, trend components, mutation points and state transition characteristics; fluctuation characteristic analysis is based on the time series pattern, and uses fluctuation spectrum analysis to quantitatively evaluate the dynamic characteristics of the risk marker value, such as the change amplitude, change rate, fluctuation frequency and stability; by integrating the time series pattern information and fluctuation characteristic evaluation results, and using principal component analysis and clustering algorithms to perform feature dimensionality reduction and pattern classification, a risk fluctuation characteristic description that can be used to reflect the temporal evolution law of physiological state risk is finally formed;

[0176] Based on the description of risk fluctuation characteristics, stability assessment and outlier detection are performed to obtain a physiological risk score. Stability assessment refers to the statistical analysis of the time series of physiological state risk marker values ​​to obtain the variance, coefficient of variation, and fluctuation frequency of the corresponding physiological state risk marker values, and based on this, a stability index and fluctuation trend chart are generated. Outlier detection uses a time series anomaly detection algorithm and combines it with the fluctuation trend chart to identify peaks, mutation points, and intervals of continuous deviation from the baseline in the time series of physiological state risk marker values, and classifies and divides them into risk levels.

[0177] Extract features from the obtained historical angiography examination records to obtain a corresponding historical angiography feature set, which includes contrast agent dosage, adverse reaction records, and examination completion quality;

[0178] Contrast tolerance analysis and sensitivity prediction are performed based on the historical angiography feature set to obtain a historical risk score. Contrast tolerance analysis refers to mining the dose-response relationship pattern in the patient's historical angiography examination records and building an individualized tolerance model based on it. The individualized tolerance model integrates key features such as previous contrast agent dosage, adverse reaction type and severity, time of occurrence and duration. Sensitivity prediction is based on a time series analysis of historical reaction data based on a machine learning algorithm to identify potential sensitivity trend changes, and combines the Bayesian network to predict the risk probability distribution corresponding to the corresponding angiography process. The historical risk score is generated by comparing the predicted risk probability distribution with the clinical preset threshold. The historical risk score can be used to reflect the patient's individual specific risk level to contrast agents inferred based on historical angiography examination records.

[0179] Establishing a three-source data feature space involves extracting features and performing dimensionality reduction on the stress safety score, physiological risk score, and historical risk score to obtain a fused feature vector. Establishing the three-source data feature space involves constructing a high-dimensional feature representation framework encompassing stress, physiological, and historical dimensions. This process utilizes tensor decomposition technology and multi-view learning methods.

[0180] A multi-level security assessment model is constructed based on the fused feature vectors to obtain preliminary security assessment results. The multi-level security assessment model construction refers to the design of a hierarchical security risk assessment architecture based on the fused feature vectors. This construction process is achieved through the use of machine learning technologies such as deep neural networks, decision tree ensembles, and probabilistic graphical models.

[0181] The obtained fusion feature vector is input into the constructed multi-level safety assessment model to generate a multi-dimensional preliminary safety assessment result including safety level, confidence interval and risk distribution;

[0182] A transfer learning method was used, and a clinical expert knowledge base was introduced to calibrate and optimize the preliminary safety assessment results to obtain the angiography safety corresponding to the corresponding vascular sub-region of the template patient.

[0183] It should be further explained that, in the specific implementation process, the process of obtaining the optimal parameter path for contrast agent injection includes:

[0184] Obtain the angiography safety data and pre-set imaging quality requirement parameters for each vascular sub-region, and construct corresponding optimization target constraints based on them. The imaging quality requirement parameters include contrast requirements, spatial resolution requirements, and temporal resolution requirements. The optimization target constraints define the objective function and constraints for the optimization of angiography injection parameters, including safety constraints and quality requirement constraints.

[0185] Based on the optimization objective constraints, a mapping relationship between injection parameters, safety, and imaging quality is established, and a corresponding parameter-effect response surface is constructed based on this relationship. The parameter-effect response surface can be used to reflect the comprehensive impact of injection parameter adjustment on safety and imaging quality.

[0186] Constructing a corresponding parameter state space based on the contrast injection parameters corresponding to the corresponding contrast injection process, wherein the parameter state space includes all possible parameter combinations and their evolution paths during several contrast injection processes; the contrast injection parameters include injection rate, pressure, total amount, time, pulse mode, etc.;

[0187] An action set and state transition function are defined in parameter space to obtain a Markov decision process model for injection control. The action set includes basic operations such as increasing / decreasing rate and increasing / decreasing pressure. The action set definition precisely quantifies the operations that can be performed by the high-pressure injection device into discrete actions, including basic operations such as increasing / decreasing rate and increasing / decreasing pressure, to form a complete control instruction library. The state transition function establishes the probability mapping relationship between the high-pressure injection device and the next state after executing a control action from the current parameter state. The state evolution equations are then obtained by fitting fluid mechanics and historical patient physiological data.

[0188] Then, based on the parameter state space, action set and state transition function, a complete model of Markov decision process represented by the four-tuple (S, A, P, R) is constructed;

[0189] A deep reinforcement learning reward function is designed by comprehensively considering factors such as angiography safety, imaging quality satisfaction, contrast agent dosage, and operation time. Based on this reward function, a deep Q-network or policy gradient network is trained to obtain a corresponding injection strategy model. The injection strategy model is adjusted and optimized in real time through online learning methods to obtain an adaptive injection control strategy. Online learning can continuously optimize the strategy based on real-time feedback, allowing the high-pressure injection device to adapt to individual patient differences and environmental changes.

[0190] Based on the adaptive injection control strategy, path planning and trajectory optimization are performed in the parameter state space to obtain an injection parameter time series. Path planning refers to searching for the optimal parameter change path in the parameter state space to ensure that contrast agent injection meets both imaging requirements and patient safety. The trajectory optimization process considers key factors such as the dynamic changes in vascular pressure, the patient's real-time physiological state, the contrast agent distribution dynamics, and the anatomical characteristics of the target vessel, and combines the injection strategy model to predict the optimal injection parameter settings at each time point, ultimately generating an injection parameter time series that includes parameters such as injection rate, pressure, and pulse mode that change precisely over time.

[0191] The injection parameter time series is safety verified and robustly tested to obtain the optimal parameter path for the corresponding contrast agent injection; based on this, the corresponding injection control instructions are generated and transmitted to the high-pressure injection equipment for execution.

[0192] In the embodiment of the present application, through the synergistic effect of multiple modules and the application of deep learning algorithms, precise and personalized control of contrast agent injection parameters is achieved, which not only significantly improves the safety of angiography but also ensures the effectiveness of imaging quality. The patient's multi-dimensional physiological data and historical examination records are obtained through the API interface, and a comprehensive analysis is performed in combination with the contrast agent characteristics to achieve accurate calculation of individualized injection parameter coefficients, which takes into account both the patient's physical differences and the organ function status; through the acquisition of vascular anatomical structure information and fluid dynamics analysis, the tolerance and pressure sensitivity of each vascular sub-region can be accurately assessed, providing a precise anatomical basis for safe injection. Based on real-time physiological monitoring data, Based on the constructed physiological status assessment model, a dynamic risk marking and early warning mechanism is realized, which improves the inspection efficiency while ensuring the safety of patients. Through the multi-source data fusion of injection pressure adjustment factors, physiological status risk marking values ​​and historical angiography records, an angiography safety degree that comprehensively reflects the safety status of blood vessels is constructed, providing a reliable decision-making basis for injection control. With the help of deep reinforcement learning algorithm, the injection parameter path is optimized, and real-time dynamic adjustment and adaptive control of parameters are realized, which not only maximizes the imaging effect but also minimizes the safety risk. Through continuous evaluation and parameter optimization mechanism, it can continuously adapt to the individual differences of patients and clinical examination needs, providing strong support for precision medicine.

[0193] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will be able to modify the technical solutions described in the foregoing embodiments or to substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0194] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0195] In the description of the present invention, it should be understood that the terms "first", "second", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0196] In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0197] In the description of the present invention, “several” means one or more, and “a large number” means two or more.

[0198] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0199] The formulas in this manual are all dimensionless and calculated using numerical values. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field based on actual conditions.

[0200] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A high-pressure injection system for angiography, characterized in that: include: The data acquisition module is used to collect the physiological data and medical records of the target patient to obtain the corresponding patient physiological data information; The patient physiological data information includes standardized patient information, physiological monitoring data, renal function assessment data and historical angiography examination records; The parameter calculation module obtains the corresponding individualized contrast agent injection parameter coefficient based on standardized patient information and renal function assessment data and combined with the contrast agent type; The vascular analysis module is used to obtain the vascular anatomical structure information corresponding to the target patient's vascular region, perform pressure assessment based on hemodynamic characteristics, and obtain the vascular pressure capacity corresponding to each vascular sub-region within the corresponding vascular region; and combine the patient's physiological data information to obtain the injection pressure adjustment factor corresponding to each vascular sub-region; The process of obtaining the vascular anatomical structure information corresponding to the target patient's vascular region includes: Performing initial impact acquisition of the patient's vascular region based on a pre-selected angiography device to obtain corresponding original vascular images; Performing digital enhancement and noise suppression on the original vascular images to obtain corresponding optimized vascular images; Performing edge detection and contour extraction on the obtained optimized blood vessel image to obtain corresponding blood vessel contour data; Obtaining the corresponding vascular diameter of each blood vessel in the target vascular region based on the extracted vascular contour data; and generating a corresponding vascular diameter distribution map based thereon; Performing three-dimensional reconstruction based on the obtained blood vessel diameter distribution map and the blood vessel diameter distribution map to obtain a spatial structure model corresponding to the corresponding blood vessel region; Based on the spatial structure model, the curvature is quantitatively calculated and the bending points are identified to obtain the corresponding vascular curvature index set; By applying a topological analysis algorithm, branch points within the corresponding spatial structure model are identified and the angles between the branch vessels and the main vessels at the corresponding branch points are measured. At the same time, the three-dimensional coordinates of the corresponding branch points are recorded. Based on the three-dimensional coordinates and angles of the branch points, the corresponding vascular branch topology is constructed. Based on the obtained vascular branch topology and the supply tissue range of the branch vessels corresponding to the branch points, the importance of the branch vessels is evaluated, and the blood flow ratio in the corresponding branch vessels is estimated by applying a pre-built fluid dynamics model. The obtained blood flow ratio and importance evaluation results are summarized to obtain the corresponding branch characteristic data; Identifying and extracting anatomical feature points corresponding to the target vascular region based on the obtained vascular diameter distribution map, vascular tortuosity index set, and branch characteristic data; and dividing the corresponding vascular region into a plurality of continuous vascular sub-regions based on the anatomical feature points; Establish an anatomical index structure for the corresponding vascular sub-region to form complete vascular anatomical structure information; The physiological assessment module assesses the physiological status of the target patient's vascular area based on the patient's physiological data information, obtains the corresponding physiological status influencing factors, and combines the pre-established physiological status safety threshold to perform risk marking and obtain the corresponding physiological status risk marking values ​​at different times; A safety construction module is used to fuse multi-source data based on the obtained injection pressure adjustment factor, physiological state risk marker value and historical angiography examination records, and to construct the angiography safety degree corresponding to the corresponding vascular sub-region based on the obtained data; The injection control module is used to obtain the optimal parameter path for contrast agent injection corresponding to the target patient based on the obtained contrast safety and combined with the deep reinforcement learning algorithm, and control the high-pressure injection equipment to perform the injection operation based on it.

2. The angiography high-pressure injection system according to claim 1, characterized in that: The process of obtaining patient physiological data information includes: Setting up a data collection unit, crawling data based on the data collection unit, and obtaining corresponding basic archival information and historical angiography examination records; Digitally process and standardize the storage of collected basic archival information to obtain corresponding standardized patient information; A data monitoring unit is provided, and physiological parameters of the patient are collected in real time based on the data monitoring unit; corresponding physiological parameter data is obtained; and data filtering and noise elimination processing are performed on the collected physiological parameters to obtain smooth physiological monitoring data; The data monitoring unit is also used to collect data on the patient's renal function-related indicators to obtain corresponding renal function assessment data; The obtained standardized patient information, physiological monitoring data, renal function assessment data and historical angiography examination records are summarized to obtain the corresponding patient physiological data information.

3. The angiography high-pressure injection system according to claim 2, characterized in that: The process of obtaining the individualized contrast agent injection parameter coefficients includes: Obtaining a physical health index corresponding to the target patient based on the standardized patient information; and obtaining a weight range corresponding to the target patient based on the physical health index; Perform benchmark dose mapping based on the target patient's weight range to obtain a benchmark dose value corresponding to the target patient's weight; and perform unit weight injection volume adjustment and upper limit control on the obtained benchmark dose value to obtain a corresponding weight adjustment coefficient; Obtain age grouping results corresponding to the corresponding patients based on standardized patient information; and perform physiological function attenuation estimation and compensation factor calculation based on the age grouping results to obtain the corresponding age correction coefficient; The collected renal function assessment data are classified into renal function grades and risk assessments to obtain the corresponding renal function status index; Based on the obtained renal function status index, a corresponding renal load assessment model is constructed, and the renal function risk level of the corresponding patient is divided into risk levels based on the renal function assessment data to obtain the corresponding renal function risk level; Predict contrast agent clearance and set safety thresholds based on the patient's corresponding renal function risk level; obtain the corresponding renal function protection coefficient; Obtain contrast agent type information; perform fluid dynamics analysis and injection characteristic simulation based on the contrast agent type information to obtain corresponding contrast agent characteristic correction factors; The weight adjustment coefficient, age correction coefficient, renal function protection coefficient and contrast agent characteristic correction factor are used as input variables; and weights are assigned to each input variable. By applying a weighted fusion algorithm, each input variable is integrated to obtain the corresponding individualized contrast agent injection parameter coefficient.

4. The angiography high-pressure injection system according to claim 1, characterized in that: The acquisition of the vascular pressure bearing capacity corresponding to each vascular sub-region includes: The blood velocity and blood flow volume corresponding to each vascular sub-region are measured in combination with the obtained anatomical structure information; a three-dimensional reconstruction is performed based on the measurement results to obtain a corresponding blood velocity vector field; and the Reynolds number corresponding to each sub-region is obtained based on the obtained blood velocity vector field; based on the obtained blood velocity vector field, a corresponding hemodynamic characteristic index set is obtained; Identify the corresponding optimized vascular images, obtain the vascular wall thickness in each vascular sub-region, and construct a corresponding wall thickness distribution map based on it; Statistical analysis is performed on the obtained wall thickness distribution map to obtain corresponding statistical feature information; anomaly detection algorithms are used to identify abnormal points in the wall thickness; corresponding vascular wall structural features are obtained; internal pressure loads are applied based on the vascular wall structural features to obtain corresponding stress-strain curves; and wall tension is calculated based on the stress-strain curves to obtain corresponding elastic modulus distributions; The maximum tolerable pressure is estimated based on the obtained elastic modulus distribution to obtain the corresponding theoretical upper limit of pressure; Obtain target patient's age, vascular disease history and other vascular health data based on standardized patient information; perform vascular wall strength correction and safety factor setting based on vascular health data; and obtain corresponding vascular status adjustment factors; The vascular pressure bearing capacity of each vascular sub-region is obtained based on hemodynamic characteristic indicators, vascular wall structural characteristics, theoretical pressure upper limit and vascular state adjustment factor.

5. The angiography high-pressure injection system according to claim 4, characterized in that: The process of obtaining the injection pressure adjustment factor includes: Based on the collected physiological monitoring data, the heart rate change curve corresponding to the target patient is obtained, and the heart rate variability analysis and periodic feature extraction are performed to obtain the corresponding heart rate fluctuation pattern; Based on the heart rate fluctuation pattern, the cardiac systolic and diastolic cycles are divided and the time window is identified to obtain the optimal injection timing index; Extracting statistical features from the systolic and diastolic blood pressure data of the corresponding patient based on the physiological monitoring data to obtain corresponding blood pressure state features; Obtain the vascular pressure bearing capacity of each vascular sub-region and construct a pressure safety threshold curve based on it; set the injection pressure upper limit and risk level based on the pressure safety threshold curve to obtain a sub-region pressure limit table; Construct a three-dimensional relationship model of heart rate, blood pressure, and pressure bearing capacity, and conduct dynamic simulation analysis and sensitivity testing based on it to obtain a parameter sensitivity matrix; Optimize the pressure regulation strategy and design a smooth transition based on the parameter sensitivity matrix to obtain the corresponding pressure regulation benchmark value; The corresponding injection pressure adjustment factor is generated according to the optimal injection timing index, the internal pressure fluctuation range, the sub-area pressure limit table and the pressure adjustment reference value.

6. The angiography high-pressure injection system according to claim 5, characterized in that: The process of obtaining physiological status influencing factors includes: Extract features from the collected physiological monitoring data to obtain the corresponding breathing pattern features of the corresponding patients; Respiratory cycle identification and intrathoracic pressure change estimation are performed based on respiratory pattern characteristics to obtain the respiratory influence coefficient; Based on the collected physiological monitoring data, the target patient's ECG waveform is obtained, and waveform features and abnormal pattern recognition are performed to obtain the corresponding heart rhythm state assessment results. Based on the heart rhythm state assessment results, cardiac output changes are predicted and hemodynamic impact analysis is performed to obtain the heart function impact coefficient; Perform trend analysis and threshold monitoring on changes in blood oxygen saturation within physiological monitoring data to obtain corresponding oxygenation status data; perform tissue perfusion assessment and hypoxia risk prediction based on oxygenation status data to obtain the oxygenation impact coefficient; Based on the obtained vascular anatomical structure information, a regional feature description corresponding to the target vascular sub-region is obtained; based on the regional feature description, risk weight allocation and key point identification are performed to obtain a regional sensitivity coefficient; The physiological state influence factors at each moment are obtained by integrating the respiratory influence coefficient, cardiac function influence coefficient, oxygenation influence coefficient and regional sensitivity coefficient.

7. The angiography high-pressure injection system according to claim 6, characterized in that: The process of obtaining the physiological status risk marker value includes: Construct safety reference standards; make individual adjustments and special case corrections to the obtained safety reference standards to obtain the corresponding personalized safety threshold set; Obtain the corresponding physiological state influencing factors at each moment, and perform component decomposition and parameter mapping based on them to obtain a multidimensional physiological index vector; Compare the multidimensional physiological indicator vectors with the personalized safety threshold set one by one and calculate the deviation to obtain a parameter deviation list; Quantify the risk level and assign weights based on the parameter deviation list to obtain a weighted risk score; Risk classification and criticality identification are performed based on weighted risk scores to obtain risk classification results; Based on the time series of the influencing factors of the current physiological state and the corresponding physiological monitoring parameters, time series risk prediction and trend analysis are performed to obtain a risk evolution trend chart; Based on the risk evolution trend chart, the warning level is determined and the risk cumulative effect is evaluated to obtain a dynamic risk index; The risk grading results and dynamic risk index are integrated, and the comprehensive risk assessment results are generated through fuzzy logic reasoning algorithm, which are then clarified to obtain the corresponding physiological status risk marker value.

8. The angiography high-pressure injection system according to claim 7, characterized in that: The process of obtaining angiographic safety includes: Obtain the injection pressure regulation factor corresponding to each vascular sub-region, and perform spatial distribution analysis and key area identification based on it to obtain a pressure-sensitive area map; Based on the pressure sensitive area map, risk level assessment and safety boundary delineation are performed to obtain a pressure safety score; Organize the physiological state risk marker values ​​at each moment into a time series, and conduct time series pattern mining and fluctuation characteristic analysis to obtain the risk fluctuation characteristics; Based on the description of risk fluctuation characteristics, stability assessment and abnormal point detection are performed to obtain physiological risk scores; Feature extraction is performed on the obtained historical angiography examination records to obtain the corresponding historical angiography feature set; based on the historical angiography feature set, angiography tolerance analysis and sensitivity prediction are performed to obtain a historical risk score; Establish a three-source data feature space, perform feature extraction and dimensionality reduction on the stress safety score, physiological risk score, and historical risk score to obtain a fused feature vector. Establishing the three-source data feature space refers to constructing a high-dimensional feature representation framework encompassing the three dimensions of stress, physiology, and history. A multi-level safety assessment model is constructed based on the fusion feature vector to obtain preliminary safety assessment results; The obtained fusion feature vector is input into the constructed multi-level security assessment model to generate a multi-dimensional preliminary security assessment result; The preliminary safety assessment results are calibrated and optimized to obtain the angiographic safety of the template patient in the corresponding vascular sub-region.

9. The angiography high-pressure injection system according to claim 8, characterized in that: The process of obtaining the optimal parameter path for contrast agent injection includes: Obtain angiography safety data and pre-set imaging quality requirement parameters for each vascular sub-region, and construct corresponding optimization target constraint conditions based on them; Based on the optimization target constraints, the mapping relationship between injection parameters, safety and imaging quality is established, and the corresponding parameter-effect response surface is constructed based on it; Constructing a corresponding parameter state space based on contrast injection parameters corresponding to the corresponding contrast injection process; Define the action set and state transition function in the parameter space to obtain the injection control Markov decision process model; Based on the parameter state space, action set and state transition function, a complete model of Markov decision process represented by four tuples is constructed; Design a deep reinforcement learning reward function and train a deep Q network or policy gradient network based on it to obtain the corresponding injection strategy model; use online learning methods to adjust and optimize the injection strategy model in real time to obtain an adaptive injection control strategy; Based on the adaptive injection control strategy, path planning and trajectory optimization are performed in the parameter state space to obtain the injection parameter time series; The injection parameter time series is subjected to safety verification and robustness testing to obtain the optimal parameter path for the corresponding contrast agent injection; based on this, the corresponding injection control instructions are generated and transmitted to the high-pressure injection equipment for execution.

Images