CN-121971737-A - Intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent from extravasation

CN121971737ACN 121971737 ACN121971737 ACN 121971737ACN-121971737-A

Abstract

The invention relates to the technical field of medical instruments, and discloses a high-pressure injection intelligent sensing and closed-loop control system for preventing contrast agent from extravasation, which is characterized in that a patient-specific vascular baseline model is established through pre-injection of normal saline, the accurate adaptation of an individualized safety threshold is realized, the hydraulic pressure and the skin tissue tension of a pipeline are synchronously acquired through a double-node pressure sensor array, the extravasation and the non-extravasation are accurately distinguished by combining a multi-feature fusion algorithm, the crossing from the passive triggering after the post to the active intervention before the pre-event is realized through a four-stage risk grading prejudgment mechanism, when the highest risk level is triggered, the system synchronously starts a dual redundancy executing mechanism, the contrast agent is guided to a waste liquid cavity through a three-way valve to realize reliable blocking through the synergism of the physical clamping pipeline of an electric pinch valve, and simultaneously, the normal saline is continuously infused in a 1-3ml/h mode is automatically started, the smoothness of a venous access is maintained at the moment of cutting off the contrast agent infusion, and the secondary risk of the blocking of the access after the cutting off of the prior art is thoroughly solved.

Inventors

MENG LIN
JIN JUAN
CHEN MENG

Assignees

连云港市第一人民医院

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. The intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent from extravasation is characterized by comprising a high-pressure injection pump execution unit and a system database, and further comprising: The pre-injection baseline acquisition and blood vessel characteristic intelligent evaluation module is used for constructing a patient-specific blood vessel baseline model through trial injection and generating a personalized safety threshold; the double-node multidimensional intelligent sensing and data preprocessing module is used for synchronously collecting the pipeline hydraulic pressure and the skin tissue tension of the puncture part and carrying out data fusion and feature extraction; The multi-feature fusion intelligent decision and risk classification pre-judging module is used for dividing a risk state into four grades of normal state, primary early warning, secondary early warning and tertiary response based on a blood vessel baseline model and real-time feature data, and generating a corresponding classification intervention instruction; The multi-feature fusion intelligent decision and risk classification prejudging module comprises a step type intervention instruction generating unit, a first-stage early warning device and a second-stage early warning device, wherein the step type intervention instruction generating unit is used for generating corresponding differential intervention instructions based on risk grades, wherein the normal state is used for maintaining injection, the first-stage early warning device is used for reducing injection flow rate and adjusting double-flow mixed injection proportion of contrast agent and normal saline, the second-stage early warning device is used for suspending injection of the contrast agent and switching the contrast agent into a KVO vein patency maintaining mode, and the third-stage response triggers blocking of a double redundant injection passage and the highest-stage early warning; the stepped redundant execution and venous access closed-loop protection module is used for executing the hierarchical intervention instruction and feeding back the execution effect; The stepwise redundant execution and venous access closed loop protection module comprises: the command matching unit is used for receiving the grading intervention command and is linked with the high-pressure injection pump executing unit, the three-way valve executing mechanism, the electric pinch valve and the KVO infusion unit; The grading intervention execution unit synchronously starts the dual redundancy execution mechanism aiming at the three-stage response instruction, drives the three-way valve to guide the contrast agent to the sealed waste liquid cavity through the servo motor, and simultaneously drives the electric pinch valve to pinch the injection pipeline to realize physical dual blocking; The multi-scene self-adaptive communication and grading early warning management and control module is used for adaptively transmitting early warning signals based on risk grades and on-site signal strength; and the system self-learning and database iterative optimization module is used for optimizing model parameters and updating a database based on clinical data.
2. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 1, wherein said intelligent assessment module for pre-injection baseline acquisition and vascular characteristics comprises: A test injection control unit for controlling the high-pressure injection pump execution unit to inject 2-5ml of physiological saline into the venous access of the patient at a flow rate of 0.5-1 ml/s; The blood vessel characteristic acquisition and analysis unit is used for acquiring a pressure response curve, a pressure change rate, a pipeline-skin pressure difference characteristic and a pressure fluctuation attenuation coefficient in the test injection process, and quantitatively evaluating the elasticity, brittleness and compliance of the blood vessel after wavelet transformation denoising; The baseline model construction and risk identification unit is used for automatically identifying extravasation high-risk people and generating an exclusive blood vessel baseline model based on the processed characteristic parameters by comparing the high-risk blood vessel characteristic library through the support vector machine classifier; And the personalized threshold generation unit is used for generating an adaptive personalized safety pressure threshold and risk judgment standard based on the blood vessel baseline model and the risk identification result.
3. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 1, wherein said dual-node multi-dimensional intelligent sensing and data preprocessing module comprises: the double-node sensing acquisition unit synchronously acquires pipeline hydraulic pressure and skin tissue tension data at millisecond sampling frequency through a medical pressure sensor arranged at the puncture end of the injection pipeline and a flexible array pressure sensor attached to the skin of the puncture part; the edge calculation preprocessing unit is used for carrying out Kalman filtering, self-adaptive noise reduction and time domain synchronous alignment and baseline calibration based on Beidou time service on the two paths of data; the two-dimensional differential fusion analysis unit is used for calculating the cross correlation coefficient and the pressure gradient difference of two paths of data and completing abnormal primary screening through the data change synchronism, the amplitude difference and the trend matching degree; the core feature extraction unit is used for extracting pressure absolute value, pressure change rate, pressure change acceleration, two paths of data differential offset, sample entropy of the pressure signal and approximate entropy features from the fusion data.
4. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 1, wherein said multi-feature fusion intelligent decision and risk classification pre-judgment module further comprises: The reference data matching unit is used for calling the personalized safety threshold value and matching real-time characteristics with the blood vessel baseline model; The abnormal cause identification unit is internally provided with a multi-feature fusion identification model based on fusion of a gradient lifting decision tree and a long-term and short-term memory network, and is used for accurately distinguishing pressure abnormality caused by extravasation and pipeline bending, body position movement, vasospasm and needle attachment by matching an extravasation feature fingerprint library with a non-extravasation abnormal feature fingerprint library; And the risk classification judging unit takes the pressure change rate, the change acceleration, the baseline deviation and the two-way data cross-correlation attenuation coefficient as core indexes, and performs the risk state classification before the pressure does not reach the safety threshold.
5. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 1, wherein said stepped redundancy execution and venous access closed-loop protection module further comprises: The venous access protection unit is used for starting the KVO infusion unit to continuously infuse physiological saline at the flow rate of 1-3ml/h while triggering the injection cutting-off action so as to avoid access blockage caused by blood reflux coagulation; And the execution effect checking unit is used for acquiring pressure data in real time through the sensing module, calculating deviation between the falling rate and a target threshold value, automatically completing three-way valve resetting and injection passage recovery when the pressure falls back to a safety threshold value, and reversely transmitting execution feedback data to the sensing module and the decision module to form a closed loop.
6. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 1, wherein said multi-scenario adaptive communication and hierarchical early warning management and control module comprises: the signal detection and region classification unit divides the operation region into high, medium and low signal intensity regions based on the received signal intensity indication value and the signal to noise ratio; The early warning grading unit divides the early warning information into three grades of mild prompt, moderate early warning and emergency warning; The self-adaptive communication adapting unit is used for adopting 5G communication and adjusting transmission power in a high signal strength area, adopting 4G/5G cellular mobile communication and adjusting the direction of a cellular antenna in a medium signal strength area, adopting Beidou short message satellite communication and adjusting modulation parameters in a low signal strength area; And the early warning execution unit is used for executing corresponding control actions based on the early warning level and recording the early warning whole-flow log and the treatment feedback data to the system database.
7. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 1, wherein said system self-learning and database iterative optimization module comprises: the data desensitization and feature extraction unit is used for retrieving the full-flow clinical data of injection from the system database to finish the privacy desensitization of patients and construct a standardized training data set; the model increment learning unit is used for carrying out increment learning and optimization on the blood vessel feature recognition, abnormal incentive classification and leakage risk pre-judging model based on the on-line random forest and elastic weight consolidation, and enriching a feature fingerprint library; The parameter optimization unit is used for presetting reference parameters, weight ratio and intervention threshold values based on the clinical data statistical analysis result self-adaptive optimization system; And the database updating unit synchronously updates the optimized model, parameters and feature library to the system database.
8. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media of claim 5, wherein the step redundancy execution and venous access closed-loop protection module execution effect verification unit performs the following analysis process: The method comprises the steps of acquiring skin tissue pressure data in a continuous time window after cutting from a double-node multidimensional intelligent sensing and data preprocessing module in real time, calculating the falling rate of pressure change along with time, carrying out correlation analysis on the falling rate and a personalized tissue compliance coefficient generated by a pre-injection baseline acquisition and blood vessel characteristic intelligent evaluation module, simultaneously carrying out dynamic comparison on a real-time pressure value and a personalized safety threshold, and when the skin pressure value is monitored to be lower than the safety threshold in three continuous sampling periods and the falling rate of the pressure tends to be stable, judging that the extravasation risk is relieved, and automatically generating a three-way valve reset signal and an injection path recovery instruction.
9. The intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent extravasation according to claim 1, wherein the multi-feature fusion intelligent decision and risk classification pre-judging module is internally provided with a multi-parameter fusion-based contrast agent extravasation risk index calculation model, and the real-time risk index is obtained by fusion calculation according to the ratio of the absolute value of the real-time pipeline pressure to the personalized pressure safety threshold value, the ratio of the pressure change rate to the maximum safety pressure change rate, the differential offset of the pipeline pressure and the skin tissue pressure, the sample entropy of a pressure signal and the cross-correlation coefficient of the pipeline pressure and the skin tissue pressure signal, and the risk index is used for quantifying the comprehensive risk level of contrast agent extravasation.
10. The intelligent sensing and closed-loop control system for high-pressure injection for preventing extravasation of contrast media according to claim 9, wherein the multi-feature fusion intelligent decision and risk classification pre-judging module is further internally provided with a risk index-based classification intervention decision model, and the intervention priority index is calculated according to the risk index, the residual response time predicted based on the pressure change trend, the ratio of the current injection flow rate to the maximum allowable injection flow rate, and the volume of the extravased contrast media estimated based on the skin tissue pressure change and the tissue compliance coefficient, and the intervention priority index is compared with a classification threshold preset by the system, so as to respectively correspond to a normal state, a primary early warning, a secondary early warning and a tertiary response.

Description

Intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent from extravasation Technical Field The invention relates to the technical field of medical instruments, in particular to a high-pressure injection intelligent sensing and closed-loop control system for preventing contrast agent from extravasation. Background The intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent from being extravasated is an intelligent medical safety protection device integrating sensor monitoring, real-time data processing and automatic executing mechanisms, and aims to solve the problem of contrast agent extravasation which is easy to occur due to high injection flow speed and high pressure in the process of high-pressure injection of contrast agent. For the above and related art, there are often the following drawbacks: The existing contrast agent extravasation prevention technology takes a preset general fixed pressure threshold value at the equipment end as a unique judgment standard, adopts a single-node and single-dimensional pressure monitoring mode, leads to series of linkage problems due to homologous defects of bottom logic, cannot adapt to individual characteristics of patients with extremely large differences of vascular physiological conditions such as old people, infants, chemotherapy and the like, cannot capture weak pressure changes in early stage of latent extravasation, causes high-risk crowd latent extravasation high-leakage detection, cannot distinguish pressure abnormality caused by non-extravasation factors such as extravasation, pipeline bending, body position movement, vasospasm and the like, can only realize one-cut binary judgment for judging extravasation by exceeding the threshold value, leads to high false positive misjudgment, can not realize pre-prediction of leakage risk through pressure change trend, can only passively trigger actions after leakage occurs, and has inherent response hysteresis; The existing contrast agent extravasation prevention technology can only realize single-point post-treatment of alarm-single injection cutting, the homologous defect of the underlying logic of the contrast agent extravasation prevention technology leads to series linkage problems, a multidimensional cooperative execution mechanism is not available, a matched venous access protection design is not available when contrast agent infusion is blocked, access blockage is very easy to be caused, secondary risks that patients need to puncture repeatedly are not available, a CT/MRI machine room shielding environment is not adapted, medical staff cannot monitor grading early warning and self-adaptive communication mechanisms of scenes on site, reliable transmission of full scenes of early warning signals cannot be guaranteed, and finally management and control failure dilemma that treatment reliability is insufficient, secondary risks frequently occur, and system performance cannot be optimized continuously is finally caused. Disclosure of Invention The intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent extravasation is provided for solving the technical problems that in the prior art, the linkage type risk identification is invalid due to incapability of adapting to clinical individual differences and complex scenes, the treatment invalidation and secondary clinical risks are caused due to no full-flow collaborative management and continuous optimization capability. In order to achieve the aim, the application adopts the following technical scheme that the intelligent sensing and closed-loop control system for high-pressure injection for preventing contrast agent from extravasation comprises a high-pressure injection pump execution unit and a system database, and is characterized by further comprising: The pre-injection baseline acquisition and blood vessel characteristic intelligent evaluation module is used for constructing a patient-specific blood vessel baseline model through trial injection and generating a personalized safety threshold; the double-node multidimensional intelligent sensing and data preprocessing module is used for synchronously collecting the pipeline hydraulic pressure and the skin tissue tension of the puncture part and carrying out data fusion and feature extraction; The multi-feature fusion intelligent decision and risk classification pre-judging module is used for carrying out risk classification and generating a classified intervention instruction based on the blood vessel baseline model and the real-time feature data; the stepped redundant execution and venous access closed-loop protection module is used for executing the hierarchical intervention instruction and feeding back the execution effect; The multi-scene self-adaptive communication and grading early warning management and control module is used for adaptively transmit