Early warning and monitoring system for preventing high-pressure injection swelling in CT (Computed Tomography) radiography

Abstract

The invention discloses an early warning and monitoring system for preventing CT radiography high-pressure injection swelling, and relates to the field of safety monitoring, and the early warning and monitoring system comprises the following steps: carrying out the real-time collection of injection safety data of each CT radiography part; according to the early warning and monitoring system for preventing the CT radiography high-pressure injection swelling, real-time training and analysis can be carried out through a large number of marked normal data and suspicious data by establishing a suspicious data recognition model; the model can also detect data changes and dynamically adjust the model to improve model prediction precision, so that professional medical personnel can remotely monitor the injection part of the examined person in real time, and the professional trained medical personnel can judge risk events; and the risk of misinformation is remarkably reduced, so that the success rate of enhanced examination in the radiology department is greatly improved. And meanwhile, the use of the equipment can effectively prevent the inspection accompanying personnel from receiving unnecessary X-ray ionization radiation.

Description

Technical Field

[0001] The invention relates to a safety monitoring technology, and in particular to an early warning monitoring system for preventing swelling caused by high-pressure injection of CT angiography. Background Art

[0002] Contrast medium extravasation refers to the rupture of the injected superficial vein during CT enhanced scanning, and the contrast medium leaks out of the blood vessels and penetrates into the adjacent tissue gap, causing congestion and swelling of the corresponding tissue. Contrast medium extravasation injury is a non-infectious inflammatory injury of subcutaneous tissue. Since the contrast medium is hyperosmotic, once it enters the tissue gap, it will cause an osmotic pressure gradient, thereby significantly increasing the interstitial fluid and forming local tissue edema. The edematous tissue can compress the blood vessels, resulting in reduced blood supply to the tissue, aggravating tissue ischemia and hypoxia, and a large amount of metabolites accumulate and cannot be discharged. In severe cases, tissue necrosis may occur. The clinical manifestations vary greatly. In mild cases, the range is limited, manifested as local papule-like swelling after acupuncture; in more severe cases, it manifests as swelling, tingling, and burning pain at the puncture site, and the forearm is full due to swelling; in severe cases, swelling can involve the entire upper limb soft tissue. In addition to soft tissue swelling, blisters and skin necrosis or ulcers may occur on the skin, blood vessels and nerves may atrophy, pulses may disappear, sensory paralysis may occur, fingertips may become pale, passive activity pain may increase, and even fascial compartment syndrome may occur. Commonly used iodine contrast agents include: Omnipaque (iohexol), Uvichrom (iopromide), diatrizoate, iodized oil, etc. This is more likely to happen when infants, the elderly, and unconscious and weak patients are injected. Therefore, in many cases, the patient's family members need to wear protective clothing and enter the scanning room to observe the real-time condition of the skin at the injection site for the person being examined. If the arm swells, they can call the technician to stop the examination in time. However, since the accompanying personnel are often non-medical professionals who lack medical knowledge, their ability to judge "swelling of the arm" varies greatly, and the scan is often interrupted due to "false alarms", which leads to the failure of this examination and requires re-examination at an appropriate time, wasting the already tight and precious medical resources. Summary of the invention

[0003] The purpose of the present invention is to provide an early warning monitoring system for preventing swelling caused by high-pressure injection of CT angiography, so as to solve the above-mentioned shortcomings in the prior art.

[0004] In order to achieve the above object, the present invention provides the following technical solution: an early warning monitoring system for preventing CT angiography high-pressure injection swelling, comprising the following steps:

[0005] S1, real-time collection of injection safety data of each CT angiography site;

[0006] S2, performing data distribution processing and task scheduling on the injection safety data collected in real time in step S1;

[0007] S3, performing real-time cleaning, conversion, and calculation on the data sent in step S2;

[0008] S4, real-time analysis of data, and data access, model building, model management, and model memory monitoring;

[0009] S5, build a suspicious data identification model and classify and predict the data in real time;

[0010] S6, reporting the data predicted to be suspicious data.

[0011] An early warning monitoring system for preventing swelling caused by high-pressure injection of CT contrast imaging, comprising the following modules:

[0012] A real-time data acquisition module, which is used to collect injection safety data of each CT angiography site in real time;

[0013] A data distribution processing module, the data distribution processing module is used to perform data distribution processing and task scheduling, the data distribution processing module includes a service packaging module, a workflow construction module, a multi-granularity modeling module and a task scheduling module, the service packaging module is used to complete the adaptive packaging of heterogeneous resources based on microservices and lightweight virtualization, realize the access of resources including cloud servers and lightweight edge CT imaging parts, and the service packaging of heterogeneous sensor resources, build resource service packaging tools, focus on the interface definition and implementation of resource services, the multi-granularity modeling module is used to perform multi-level modeling of the system from multiple angles through multi-level resource views and cluster views, the workflow construction module is used for task workflow construction, task workflow management and task workflow combination nesting, and the task scheduling module is used to schedule tasks according to task workflows;

[0014] A real-time data processing module, which is used for real-time cleaning, conversion, and calculation of data, including functions such as data access, real-time processing, and log management;

[0015] A real-time data analysis module, which is used for real-time analysis of data and provides data access, model building, model management, and model memory monitoring functions;

[0016] A suspicious data identification module, which is used to build a suspicious data identification model and classify and predict data in real time;

[0017] A suspicious data reporting module is used to report data predicted to be suspicious data.

[0018] Furthermore, the specific working method of the service encapsulation module is:

[0019] A1, reading the injection safety data of each CT angiography site received by the interface;

[0020] A2, parse resource input and output, transmission protocol, and data format;

[0021] A3, select the interface adapter that meets the configuration;

[0022] A4, interface adapter connects sensor resources for data encapsulation and format conversion;

[0023] A5, generate resource service implementation class;

[0024] A7, service information registration.

[0025] Furthermore, the specific working method of the multi-granularity modeling module is:

[0026] B1, read the CT angiography site information in the database:

[0027] B2, analysis of the hierarchical relationship of CT angiography sites;

[0028] B3, build a coarse-grained model;

[0029] B4, obtain the target observation level;

[0030] B5, analysis of hierarchical relationship of fine-grained CT angiography sites;

[0031] B6, build fine-grained models;

[0032] B7, sending a command to update the status of the designated CT angiography part;

[0033] B8, read the status information of the CT angiography part recovery.

[0034] Furthermore, the specific working method of the workflow construction module is:

[0035] C1, task workflow construction, the steps of task workflow construction are:

[0036] C11, performs service discovery and reads component data serialization;

[0037] C12, parse the service interface description file;

[0038] C13, calling resource service implementation class;

[0039] C14, write component configuration information;

[0040] C15, write component connection relationship;

[0041] C16, write task information;

[0042] C17, storage process files;

[0043] C2, task workflow management, which includes process viewing, process editing, information editing, and task deletion;

[0044] C3, task workflow combination nesting:

[0045] C31, read the task workflow list;

[0046] C32, find available task workflows;

[0047] C33, parses and writes component configuration;

[0048] C34, parses workflow extensible markup and generates component data serialization files;

[0049] C35, stores relevant files and information;

[0050] C36, adding new components to the optional components.

[0051] Furthermore, the specific working method of the task scheduling module is:

[0052] D1, read task workflow;

[0053] D2, find available task workflows;

[0054] D3, reads the target workflow extensible markup file;

[0055] D4, read execution mode;

[0056] D5, write execution information;

[0057] D6, Parsing Workflow Extensible Markup;

[0058] D7, assemble parameter objects and submit them to the workflow engine;

[0059] D8, the scheduler performs container orchestration.

[0060] Furthermore, the specific working method of the suspicious data identification module to construct the suspicious data identification model is:

[0061] E1, initialize decision tree and window;

[0062] E2, data stream samples flow into the node, if the number of samples is sufficient, execute step E7, if the number of samples is insufficient, execute step E3;

[0063] E3, add data stream samples to the window;

[0064] E4, determine whether the old sample window size is larger than the target window size, if the determination result is yes, delete the old sample, if the determination result is no, return to step E3;

[0065] E5, delete old samples;

[0066] E6, when the number of samples is sufficient, execute step E10;

[0067] E7, check all split attributes;

[0068] E8, determine whether concept drift occurs, if the determination result is yes, return to step E7, if the determination result is no, execute step E10;

[0069] E9, the backup subtree replaces the original branch;

[0070] E10, update decision.

[0071] Compared with the prior art, the present invention provides an early warning monitoring system for preventing CT contrast high-pressure injection swelling. Through the nesting of task workflow combinations, the components generated by the target task workflow can be added to the optional components when the new task workflow is constructed. When the task workflow is running, the system will automatically parse such nested nodes, implicitly run a subtask process, and submit container orchestration. Through this function, the heterogeneous resource distributed management and scheduling system realizes the ability of multi-level nesting of task workflows. At the same time, by establishing a suspicious data recognition model, real-time training and analysis can be performed through a large amount of marked normal data and suspicious data. When concept drift occurs in the data stream, the model will also detect data changes and dynamically adjust the model to improve the model prediction accuracy, so that professional medical personnel can perform real-time remote monitoring of the injection site of the examinee, allowing professionally trained medical personnel to judge risk events, significantly reducing the risk of "false alarms" to greatly improve the success rate of enhanced examinations in the radiology department. At the same time, the use of the device can also effectively prevent the accompanying personnel from receiving unnecessary X-ray ionizing radiation. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0073] Figure 1 A schematic diagram of the overall method flow provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0075] See also Figure 1 , an early warning monitoring system for preventing swelling caused by high-pressure injection of CT contrast imaging, comprising the following steps:

[0076] S1, real-time collection of injection safety data of each CT angiography site;

[0077] S2, performing data distribution processing and task scheduling on the injection safety data collected in real time in step S1;

[0078] S3, performing real-time cleaning, conversion, and calculation on the data sent in step S2;

[0079] S4, real-time analysis of data, and data access, model building, model management, and model memory monitoring;

[0080] S5, build a suspicious data identification model and classify and predict the data in real time;

[0081] S6, reporting the data predicted to be suspicious data.

[0082] An early warning monitoring system for preventing swelling caused by high-pressure injection of CT contrast imaging, comprising the following modules:

[0083] A real-time data acquisition module, which is used to collect injection safety data of each CT angiography site in real time;

[0084] A data distribution processing module, the data distribution processing module is used to perform data distribution processing and task scheduling, the data distribution processing module includes a service packaging module, a workflow construction module, a multi-granularity modeling module and a task scheduling module, the service packaging module is used to complete the adaptive packaging of heterogeneous resources based on microservices and lightweight virtualization, realize the access of resources including cloud servers and lightweight edge CT imaging parts, and the service packaging of heterogeneous sensor resources, build resource service packaging tools, focus on the interface definition and implementation of resource services, the multi-granularity modeling module is used to perform multi-level modeling of the system from multiple angles through multi-level resource views and cluster views, the workflow construction module is used for task workflow construction, task workflow management and task workflow combination nesting, and the task scheduling module is used to schedule tasks according to task workflows;

[0085] A real-time data processing module, which is used for real-time cleaning, conversion, and calculation of data, including functions such as data access, real-time processing, and log management;

[0086] A real-time data analysis module, which is used for real-time analysis of data and provides data access, model building, model management, and model memory monitoring functions;

[0087] A suspicious data identification module, which is used to build a suspicious data identification model and classify and predict data in real time;

[0088] A suspicious data reporting module is used to report data predicted to be suspicious data.

[0089] In this arrangement, the injection safety information data of each CT angiography site is collected in real time through the real-time data acquisition module, data distribution processing and task scheduling are performed through the data distribution processing module, the emitted data are cleaned, converted, and calculated in real time through the real-time data processing module, the real-time data analysis module performs real-time analysis of the data, and performs data access, model construction, model management, and model memory monitoring, a suspicious data identification model is constructed through the suspicious data identification module, and data is classified and predicted in real time, and data predicted as suspicious data is reported to the back-end processing personnel through the suspicious data reporting module for suspicious prediction data processing.

[0090] The specific working method of the service encapsulation module is:

[0091] A1, reading the injection safety data of each CT angiography site received by the interface;

[0092] A2, parse resource input and output, transmission protocol, and data format;

[0093] A3, select the interface adapter that meets the configuration;

[0094] A4, interface adapter connects sensor resources for data encapsulation and format conversion;

[0095] A5, generate resource service implementation class;

[0096] A7, service information registration.

[0097] The specific working method of the multi-granularity modeling module is:

[0098] B1, read the CT angiography site information in the database:

[0099] B2, analysis of the hierarchical relationship of CT angiography sites;

[0100] B3, build a coarse-grained model;

[0101] B4, obtain the target observation level;

[0102] B5, analysis of hierarchical relationship of fine-grained CT angiography sites;

[0103] B6, build fine-grained models;

[0104] B7, sending a command to update the status of the designated CT angiography part;

[0105] B8, read the status information of the CT angiography part recovery.

[0106] In this way, the system first reads the basic CT imaging site hierarchical relationship of resources from the database and builds a coarse-grained model. Under this model, users will understand the overall operation of information infrastructure resources. Users can specify the granularity of the observation model. When there is a need for fine-grained CT imaging site observation, the system will read the resource-specified target structure CT imaging site hierarchical relationship from the database according to user needs, perform fine-grained and detailed resource CT imaging site hierarchical relationship modeling, and provide the working status and information of each independent unit.

[0107] The specific working method of the workflow building module is:

[0108] C1, task workflow construction, the steps of task workflow construction are:

[0109] C11, performs service discovery and reads component data serialization;

[0110] C12, parse the service interface description file;

[0111] C13, calling resource service implementation class;

[0112] C14, write component configuration information;

[0113] C15, write component connection relationship;

[0114] C16, write task information;

[0115] C17, storage process files;

[0116] C2, task workflow management, task workflow management includes process viewing, process editing, information editing, and task deletion;

[0117] C3, task workflow combination nesting:

[0118] C31, read the task workflow list;

[0119] C32, find available task workflows;

[0120] C33, parses and writes component configuration;

[0121] C34, parses workflow extensible markup and generates component data serialization files;

[0122] C35, stores relevant files and information;

[0123] C36, adding new components to the optional components.

[0124] In this way, through the construction of task workflows, for each service node, the system will call the services that are standardized and encapsulated from the heterogeneous resources connected to the information infrastructure according to the user configuration when executing the task. At the same time, through the combination and nesting of task workflows, the components generated by the target task workflow can be added to the optional components when building a new task workflow. When the task workflow is running, the system will automatically parse such nested nodes, implicitly run a subtask process, and submit the container orchestration. Through this function, the heterogeneous resource distributed management and scheduling system realizes the ability of multi-level nesting of task workflows.

[0125] The specific working method of the task scheduling module is:

[0126] D1, read task workflow;

[0127] D2, find available task workflows;

[0128] D3, reads the target workflow extensible markup file;

[0129] D4, read execution mode;

[0130] D5, write execution information;

[0131] D6, Parsing Workflow Extensible Markup;

[0132] D7, assemble parameter objects and submit them to the workflow engine;

[0133] D8, the scheduler performs container orchestration.

[0134] This setting allows the user to enter the task configuration and execution process interface. The system reads the executable task workflow from the cloud database and provides the read results to the user in a list. The operation bar on the right side of the list can run the task workflow. The task workflow supports direct execution and scheduled execution. When the task is running, the system parses the task workflow extensible markup file, assembles it into a parameter object and submits it to the workflow engine. The scheduler will perform resource scheduling and complete container orchestration.

[0135] The specific working method of the suspicious data identification module to build a suspicious data identification model is as follows:

[0136] E1, initialize decision tree and window;

[0137] E2, data stream samples flow into the node, if the number of samples is sufficient, execute step E7, if the number of samples is insufficient, execute step E3;

[0138] E3, add data stream samples to the window;

[0139] E4, determine whether the old sample window size is larger than the target window size, if the determination result is yes, delete the old sample, if the determination result is no, return to step E3;

[0140] E5, delete old samples;

[0141] E6, when the number of samples is sufficient, execute step E10;

[0142] E7, check all split attributes;

[0143] E8, determine whether concept drift occurs, if the determination result is yes, return to step E7, if the determination result is no, execute step E10;

[0144] E9, the backup subtree replaces the original branch;

[0145] E10, update decision.

[0146] The model is trained and analyzed in real time with a large amount of labeled normal data and suspicious data. When concept drift occurs in the data stream, the model will detect data changes and dynamically adjust the model to improve the model's prediction accuracy.

[0147] Working principle: When in use, the injection safety information data of each CT angiography site is collected in real time through the real-time data acquisition module, the data distribution processing and task scheduling are carried out through the data distribution processing module, the sent data is cleaned, converted and calculated in real time through the real-time data processing module, the real-time data analysis module analyzes the data in real time, and performs data access, model construction, model management and model memory monitoring, the suspicious data identification model is constructed through the suspicious data identification module, and the data is classified and predicted in real time, and the data predicted as suspicious data is reported to the background processing personnel through the suspicious data reporting module for suspicious prediction data processing.

[0148] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An early warning monitoring system for preventing swelling caused by high-pressure injection of CT angiography, characterized in that: The steps include: S1, real-time collection of injection safety data of each CT angiography site; S2, performing data distribution processing and task scheduling on the injection safety data collected in real time in step S1; S3, performing real-time cleaning, conversion, and calculation on the data sent in step S2; S4, real-time analysis of data, and data access, model building, model management, and model memory monitoring; S5, build a suspicious data identification model and classify and predict the data in real time; S6, reporting the data predicted to be suspicious data.

2. The early warning monitoring system for preventing CT angiography high-pressure injection swelling according to claim 1 is characterized in that: Includes the following modules: A real-time data acquisition module, which is used to collect injection safety data of each CT angiography site in real time; A data distribution processing module, the data distribution processing module is used to perform data distribution processing and task scheduling, the data distribution processing module includes a service packaging module, a workflow construction module, a multi-granularity modeling module and a task scheduling module, the service packaging module is used to complete the adaptive packaging of heterogeneous resources based on microservices and lightweight virtualization, realize the access of resources including cloud servers and lightweight edge CT imaging parts, and the service packaging of heterogeneous sensor resources, build resource service packaging tools, focus on the interface definition and implementation of resource services, the multi-granularity modeling module is used to perform multi-level modeling of the system from multiple angles through multi-level resource views and cluster views, the workflow construction module is used for task workflow construction, task workflow management and task workflow combination nesting, and the task scheduling module is used to schedule tasks according to task workflows; A real-time data processing module, which is used for real-time cleaning, conversion, and calculation of data, including functions such as data access, real-time processing, and log management; A real-time data analysis module, which is used for real-time analysis of data and provides data access, model building, model management, and model memory monitoring functions; A suspicious data identification module, which is used to build a suspicious data identification model and classify and predict data in real time; A suspicious data reporting module is used to report data predicted to be suspicious data.

3. The early warning monitoring system for preventing CT angiography high-pressure injection swelling according to claim 2 is characterized in that: The specific working method of the service encapsulation module is: A1, reading the injection safety data of each CT angiography site received by the interface; A2, parse resource input and output, transmission protocol, and data format; A3, select the interface adapter that meets the configuration; A4, interface adapter connects sensor resources for data encapsulation and format conversion; A5, generate resource service implementation class; A7, service information registration.

4. The early warning monitoring system for preventing CT angiography high-pressure injection swelling according to claim 3 is characterized in that: The specific working method of the multi-granularity modeling module is: B1, read the CT angiography site information in the database: B2, analysis of the hierarchical relationship of CT angiography sites; B3, build a coarse-grained model; B4, obtain the target observation level; B5, analysis of hierarchical relationship of fine-grained CT angiography sites; B6, build fine-grained models; B7, sending a command to update the status of the designated CT angiography part; B8, read the status information of the CT angiography part recovery.

5. The early warning monitoring system for preventing CT angiography high-pressure injection swelling according to claim 4 is characterized in that: The specific working method of the workflow construction module is: C1, task workflow construction, the steps of task workflow construction are: C11, performs service discovery and reads component data serialization; C12, parse the service interface description file; C13, calling resource service implementation class; C14, write component configuration information; C15, write component connection relationship; C16, write task information; C17, storage process files; C2, task workflow management, which includes process viewing, process editing, information editing, and task deletion; C3, task workflow combination nesting, including: C31, read the task workflow list; C32, find available task workflows; C33, parses and writes component configuration; C34, parses workflow extensible markup and generates component data serialization files; C35, stores relevant files and information; C36, adding new components to the optional components.

6. The early warning monitoring system for preventing CT angiography high-pressure injection swelling according to claim 5, characterized in that: The specific working method of the task scheduling module is: task workflow combination nesting D1, read task workflow; D2, find available task workflows; D3, reads the target workflow extensible markup file; D4, read execution mode; D5, write execution information; D6, Parsing Workflow Extensible Markup; D7, assemble parameter objects and submit them to the workflow engine; D8, the scheduler performs container orchestration.

7. The early warning monitoring system for preventing CT angiography high-pressure injection swelling according to claim 6, characterized in that: The specific working method of the suspicious data identification module to construct the suspicious data identification model is: E1, initialize decision tree and window; E2, data stream samples flow into the node, if the number of samples is sufficient, execute step E7, if the number of samples is insufficient, execute step E3; E3, add data stream samples to the window; E4, determine whether the old sample window size is larger than the target window size, if the determination result is yes, delete the old sample, if the determination result is no, return to step E3; E5, delete old samples; E6, when the number of samples is sufficient, execute step E10; E7, check all split attributes; E8, determine whether concept drift occurs, if the determination result is yes, return to step E7, if the determination result is no, execute step E10; E9, the backup subtree replaces the original branch; E10, update the decision tree.

Images