An ai technology-based coronary heart disease examination path triage system

Abstract

The application provides a coronary heart disease examination path triage system based on AI technology, and relates to the technical field of medical treatment.The coronary heart disease examination path triage system based on AI technology comprises a data access module, an AI integral calculation module, a symptom judgment module, a rule engine module, a result display and interface module and a parameter management and audit module, collects non-ECG gated chest CT image data, automatically calculates coronary artery calcification integral (CACS) by using AI, combines chest pain type judgment results and clinical adaptation evidence, constructs a binary risk stratification system, and outputs low-risk or high-risk stratification results and corresponding examination path suggestions.The application solves the defects of the prior art, such as dependence on subjective experience, low efficiency of calcification integral calculation and inaccuracy of examination path matching, realizes standardized and accurate shunt decision, greatly improves diagnosis and treatment efficiency, reduces medical risks and resource waste, and is suitable for coronary heart disease diagnosis and shunt scenes of medical institutions at all levels.

Description

Technical Field

[0001] This invention relates to the field of medical technology, specifically to a triage system for coronary heart disease examination pathways based on AI technology. Background Technology

[0002] Coronary artery disease (CAD) is a major chronic disease that threatens the health of people worldwide. Early and accurate triage of suspected CAD patients is crucial to reducing diagnosis and treatment risks and controlling medical costs. Current clinical triage mainly relies on the subjective experience of physicians, combined with the selection of examination methods based on "low-intermediate-risk" and "high-risk" classifications: CT angiography (CTA) is the first choice for low-intermediate-risk patients, while coronary angiography (CAG) is used for high-risk patients.

[0003] However, existing technologies have significant drawbacks: First, the lack of unified quantitative standards for risk stratification, which relies on doctors' subjective judgment, can easily lead to low-risk patients receiving excessive invasive CAG examinations or high-risk patients experiencing delayed diagnosis and treatment. Secondly, traditional coronary artery calcium score (CACS) calculation requires manual measurement, which takes 10-15 minutes, is inefficient and has a large error, and cannot meet the needs of emergency and large number of patients. Third, the examination pathway is not well-suited to the calcification load. When the calcification load is high (CACS > 120 points), CTA is prone to over-assessing the degree of stenosis, and the current protocol does not take into account this clinical evidence, resulting in inaccurate matching. Fourth, the lack of a standardized decision-making process that combines quantitative indicators, symptom classification, and examination matching leads to both waste of medical resources and treatment risks.

[0004] Therefore, developing a coronary heart disease diagnosis and triage system that is based on objective quantitative indicators, integrates clinical evidence, and has a high degree of automation has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a coronary heart disease triage system based on AI technology. By automatically quantifying CACS scores, objectively determining chest pain types, and combining evidence of the suitability of examination methods with AI, it achieves standardized and precise triage decisions, thus solving the problems mentioned in the background technology.

[0006] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a coronary heart disease examination pathway triage system based on AI technology, comprising a data access module, an AI score calculation module, a symptom determination module, a rule engine module, a result display and interface module, and a parameter management and auditing module. The system performs the following steps: S1. Collect non-ECG-gated chest CT image data of suspected coronary heart disease patients through the data access module and transmit them to the AI ​​analysis platform in DICOM standard format; S2. The AI ​​scoring module uses a deep learning segmentation model to first segment the coronary artery region and then filter calcified regions according to a preset threshold. It automatically identifies the anatomical region of the coronary artery and calcified lesions, calculates and outputs the total CACS score and the scores of the optional left main coronary artery (LM), left anterior descending artery (LAD), circumflex artery (LCX), and right coronary artery (RCA) branches. S3. Based on three preset core features, the symptom judgment module determines whether the patient's chest pain is typical or atypical. S4. Based on AI-CACS scores, chest pain types, and evidence of the compatibility between CACS and examination methods, the rule engine module constructs a binary risk stratification system and outputs low-risk or high-risk results. S5. Based on the risk stratification results, corresponding treatment pathway suggestions are generated. For low-risk cases, coronary CTA is recommended as the priority, while for high-risk cases, direct invasive coronary angiography (CAG) is recommended. S6. Visualize the analysis results through the results display and interface module, and write key data into the hospital's information system.

[0007] Preferably, the three core characteristics in step S3 are: retrosternal discomfort with nature and duration consistent with typical coronary heart disease pain; pain that can be induced by exertion or emotional stress; pain relief within minutes after rest and / or nitrate drug treatment; all three characteristics are consistent for typical chest pain, while only two or less are consistent for atypical chest pain.

[0008] Preferably, the binary risk stratification rule in step S4 is as follows: the low-risk stratification is defined as 0 ≤ CACS ≤ 120 points and atypical chest pain; the high-risk stratification is defined as CACS > 120 points or typical chest pain, and the CACS stratification threshold supports personalized configuration by medical institutions.

[0009] Preferably, the deep learning segmentation model of the AI ​​integral calculation module in step S2 is a convolutional neural network (CNN) model with a preset density threshold ≥130HU. The Agatston scoring algorithm is used to calculate the CACS score. The Pearson correlation coefficient between this model and the ECG-gated calcium integral scan results is ≥0.985, and the intragroup correlation coefficient (ICC) for risk stratification is ≥0.942.

[0010] Preferably, when the CTA examination in step S5 indicates vascular stenosis ≥50%, the system automatically triggers a subsequent CAG examination recommendation; when the CAG examination in step S5 indicates severe stenosis, the system provides information on the indications for interventional treatment.

[0011] Preferably, the data access module in step S1 adopts the DICOM standard interface, supports data access from multiple brands of CT equipment, and has functions for patient queue classification, diagnosis and treatment task creation, and status tracking.

[0012] Preferably, the symptom determination module provides a structured chest pain feature input interface, supporting manual input or data extraction in conjunction with an electronic medical record (EMR) system.

[0013] Preferably, the result display and interface module in step S6 automatically writes key data into the hospital's HIS / LIS / PACS system via the HL7 interface, and the visualized report includes CACS score, chest pain type, risk stratification results, examination recommendations, and evidence.

[0014] Preferably, the parameter management and auditing module provides a visual parameter configuration interface, supports adjusting the CACS stratification threshold and CTA positive threshold, and records model version iteration logs, user operation trajectories and data access records to meet medical data security compliance requirements.

[0015] Preferably, the system supports in-hospital or cloud deployment, can be expanded to access multimodal data such as electrocardiograms and blood biochemical indicators, and supports algorithm model iteration updates and clinical guideline import.

[0016] (III) Beneficial Effects This invention provides a coronary heart disease examination pathway triage system based on AI technology, which has the following beneficial effects: 1. By deeply integrating AI-automated quantitative CACS scoring, chest pain classification, and CACS-examination method suitability evidence, a two-dimensional risk stratification system is constructed, which solves the core defects of existing technologies such as subjectivity and lack of quantitative standards.

[0017] 2. The AI-CACS score has a high consistency with the gold standard (gated scanning) (Pearson correlation coefficient 0.985). The risk stratification rule is based on clinical research evidence for triage, ensuring the accuracy of examination path matching and reducing the risk of misdiagnosis / missed diagnosis.

[0018] 3. AI can automatically complete CACS scoring in just 30 seconds, which is much more efficient than manual measurement (10-15 minutes) and meets the needs of rapid triage and batch patient processing in the emergency department.

[0019] 4. Low-risk patients (low CACS score 0 ≤ CACS ≤ 120 and atypical chest pain) should be given priority for non-invasive CTA to avoid unnecessary invasive procedures; high-risk patients (high CACS score > 120 or typical chest pain) should be given direct CAG to reduce ineffective CTA examinations, reduce waste of medical resources and the economic burden on patients.

[0020] 5. Based on standard DICOM data and HL7 interface, it does not require modification of existing medical equipment and information systems, supports in-hospital or cloud deployment, and can be quickly adapted to medical institutions at all levels. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2 AI-CACS calculation flowchart; Figure 3 A flowchart for risk stratification and triage decision-making; Figure 4 This is a schematic diagram of the system interface; Figure 5 This is a schematic diagram of CTA and CAG contrast imaging in high-CACS cases. Figure 6 A schematic diagram of contrast imaging in cases with high CACS where CTA and CAG are inconsistent; Figure 7 This is a schematic diagram of CTA and CAG angiography in low-CACS cases. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1: like Figures 1-4 As shown in the figure, this embodiment of the invention provides a coronary heart disease examination pathway triage system based on AI technology, including six modules: data access module, AI score calculation module, symptom judgment module, rule engine module, result display and interface module, and parameter management and auditing module.

[0024] The system performs the following steps: S1. Collect non-ECG-gated chest CT image data of suspected coronary heart disease patients through the data access module and transmit them to the AI ​​analysis platform in DICOM standard format; S2. The AI ​​scoring module uses a deep learning segmentation model to first segment the coronary artery region and then filter calcified regions according to a preset threshold. It automatically identifies the anatomical region of the coronary artery and calcified lesions, calculates and outputs the total score of coronary artery calcification score (hereinafter referred to as CACS) and the scores of optional branches of the left main coronary artery (LM), left anterior descending artery (LAD), circumflex artery (CX), and right coronary artery (RCA). S3. Based on three preset core features, the symptom judgment module determines whether the patient's chest pain is typical or atypical. S4. Based on AI-CACS scores, chest pain types, and evidence of the compatibility between CACS and examination methods, the rule engine module constructs a binary risk stratification system and outputs low-risk or high-risk results. S5. Based on the risk stratification results, corresponding treatment pathway suggestions are generated. For low-risk cases, coronary CTA is recommended as the priority, while for high-risk cases, direct invasive coronary angiography (CAG) is recommended. S6. Visualize the analysis results through the results display and interface module, and write key data into the hospital's information system.

[0025] Specifically, the system works collaboratively through six functional modules to construct a complete technical chain of "data collection - AI calculation - symptom assessment - risk stratification - path recommendation - result feedback." The specific technical solutions include: 1. Data Access Module: Adopting the Medical Digital Imaging and Communication (DICOM) standard interface, it supports non-ECG-gated chest CT image data access from multiple brands of CT equipment, and has functions for patient queue classification, diagnosis and treatment task creation and status tracking, ensuring the compatibility, stability and security of data transmission; 2. AI Scoring Module: The core innovation lies in the use of a deep learning segmentation model. First, a convolutional neural network (CNN) is used to accurately segment the coronary artery anatomy (left main coronary artery (LM), left anterior descending artery (LAD), left circumflex artery (LCX), and right coronary artery (RCA). Then, calcified lesions are screened based on a preset density threshold (≥130 HU), and the Agatston scoring algorithm is used to automatically calculate the total CACS score and the scores of each branch vessel. This module has been validated with 1000 clinical cases, achieving a Pearson correlation coefficient of 0.985 with the ECG-gated calcification scoring results, and an ICC of 0.942 for the four risk stratifications (0, 1-100, 101-400, >400 points), ensuring scoring accuracy. 3. Symptom Determination Module: Provides a structured chest pain feature input interface, supports manual input or data extraction in conjunction with the electronic medical record (EMR) system, automatically determines the type of chest pain based on three core features (pain location and nature, triggering factors, and relief methods), and outputs "typical chest pain" or "atypical chest pain" results to avoid subjective judgment errors. 4. Rule Engine Module: This module incorporates a dual-dimensional risk stratification rule. Its core innovation lies in integrating clinical evidence on the compatibility of CACS with different examination methods. Specific rules are as follows: Low-risk stratum: Low score 0 ≤ CACS ≤ 120 (good consistency between CTA and CAG) and atypical chest pain; High-risk stratum: High score CACS > 120 (CTA is prone to over-assessment) or typical chest pain (high cardiovascular risk). This module supports personalized configuration of CACS stratification thresholds and CTA positive thresholds (default stenosis ≥ 50%), and has rule version management functionality to adapt to clinical guideline updates. Triage Recommendation Generation and Execution: Based on risk stratification results, standardized treatment pathway recommendations are output. Specifically: Low-risk stratum: Coronary CTA is prioritized, fully utilizing the non-invasive and precise advantages of CTA under low calcification load; if CTA indicates vascular stenosis ≥ 50%, subsequent CAG examination recommendations are automatically triggered; High-risk stratum: CAG examination is directly recommended, avoiding the risk of CTA over-assessment and facilitating timely interventional treatment; if CAG indicates severe stenosis, the system provides information on the indications for interventional treatment. 5. Results Display and Interface Module: Generates visual reports on clinical workstations or mobile terminals, including CACS scores (total score and branch scores), chest pain type, risk stratification results, examination recommendations and supporting evidence; at the same time, it automatically writes key data into the hospital's HIS / LIS / PACS system through the HL7 interface to support medical quality control and long-term follow-up. 6. Parameter Management and Audit Module: Provides a visual parameter configuration interface, supporting medical institutions to adjust core parameters such as CACS stratification threshold and CTA positive threshold; records model version iteration logs, user operation trajectories and data access records, meeting compliance requirements such as the "Guideline for Medical Data Security".

[0026] The relevant terms in the context are explained as follows: AI-CACS: Acquires non-ECG-gated chest CT image data and automatically calculates the coronary artery calcium score using a deep learning segmentation model; Typical chest pain: Chest pain that simultaneously meets the three characteristics of "retrosternal discomfort with typical features, exertion / emotional triggering, and relief by rest / nitrates"; Atypical chest pain: Chest pain that only meets two or fewer of the above characteristics; Low-risk patients have low scores (0≤CACS≤120 points) and atypical chest pain; High-risk patients have high scores (CACS>120 points) or typical chest pain; CTA: Coronary CT angiography, a non-invasive examination method, with good consistency with CAG assessment when the calcification burden is low (0≤CACS≤120 points); CAG: Invasive coronary angiography, the gold standard for diagnosing coronary heart disease, which can avoid over-assessment by CTA when the calcification burden is high (CACS>120 points); Agatston score: An internationally recognized quantitative standard for coronary artery calcium, which calculates the score based on the density and area of ​​calcified lesions.

[0027] As an improvement, the three core characteristics in step S3 are: retrosternal discomfort with nature and duration consistent with typical coronary heart disease pain; pain that can be induced by exertion or emotional stress; pain relief within minutes after rest and / or nitrate treatment; all three characteristics must be present for typical chest pain, while only two or fewer characteristics must be present for atypical chest pain.

[0028] As an improvement, the binary risk stratification rule in step S4 is as follows: the low-risk stratification is defined as 0 ≤ CACS ≤ 120 points and atypical chest pain; the high-risk stratification is defined as CACS > 120 points or typical chest pain, and the CACS stratification threshold supports personalized configuration by medical institutions.

[0029] As an improvement, the deep learning segmentation model of the AI ​​integral calculation module has been validated with large sample data, and the Pearson correlation coefficient with the ECG-gated calcium integral scan results is ≥0.985, and the intragroup correlation coefficient (ICC) for risk stratification is ≥0.942.

[0030] As an improvement, when CTA examination indicates ≥50% vascular stenosis, CAG examination is further recommended; when CAG examination indicates severe vascular stenosis, the information level suggests an indication for interventional treatment.

[0031] As an improvement, the triage system supports in-hospital or cloud deployment. The parameter management and auditing module records model version iteration logs and user operation trajectories to meet medical data security and compliance requirements.

[0032] Example 2: Based on Example 1, this embodiment conducts the following experiments: Appendix Figure 5 In the image, Figure 1A shows coronary artery calcification on chest CT scan; Figure 1B shows an AI analysis score of 1136 for the left anterior descending artery calcification; Figure 1C shows severe stenosis with calcified plaques in the mid-segment of the left anterior descending artery on coronary CTA; and Figure 1D shows 95%-99% stenosis in the mid-segment of the left anterior descending artery confirmed by coronary angiography.

[0033] Appendix Figure 6 In the middle, Figure 2A: Chest CT scan shows coronary artery calcification; Figure 2B: AI analysis shows a calcification score of 1392 in the left anterior descending artery; Figure 2C: Coronary CTA shows severe stenosis with calcified plaques in the proximal and mid segments of the left anterior descending artery; Figure 2D: Coronary angiography confirms approximately 45% stenosis in the mid segment of the left anterior descending artery.

[0034] Appendix Figure 7 In the first image, a chest CT scan showed punctate calcifications in the coronary arteries; in the second image, AI analysis showed a calcification score of 4 in the left anterior descending artery; in the third image, coronary CTA showed severe stenosis in the proximal-mid segment of the left anterior descending artery; and in the fourth image, coronary angiography confirmed approximately 95% stenosis in the mid-segment of the left anterior descending artery.

[0035] As shown in the figure above, the coronary heart disease examination pathway triage system based on AI technology involved in this invention requires a basic hardware and software environment including CPU: Intel Xeon Gold 6248, GPU: NVIDIA Tesla V100, 64GB of memory, and the software system supports Windows Server 2016 and above operating systems. It connects to the radiology CT equipment through a DICOM gateway and links to the HIS / EMR / PACS system through an HL7 interface.

[0036] During the data acquisition process, after a patient with suspected coronary heart disease completes a non-ECG-gated chest CT scan, the image data is automatically uploaded to the system's data access module in DICOM format. The module automatically extracts the patient's basic information, including name, ID, and examination time, and then creates a treatment task.

[0037] In the AI-CACS calculation process, the AI ​​score calculation module calls a deep learning model to first segment the four major branches of the coronary artery with a segmentation accuracy of ≥95%. Then, calcified lesions with a standard density of ≥130HU are selected, and the total CACS score and branch scores are calculated. The results are then output to the rule engine module.

[0038] Next, the chest pain type is determined. The doctor enters the patient's chest pain characteristics through the system interface. This step may be automatically synchronized by the EMR system. The symptom determination module automatically outputs the result of "typical chest pain" or "atypical chest pain" according to preset rules. In this step, risk stratification and path recommendation are performed. The rule engine module combines the CACS score and chest pain type to output low-risk / high-risk stratification results and automatically generate corresponding examination suggestions. Then, the system displays a visual report on the clinical workstation. Doctors refer to the suggestions to formulate treatment plans. Key data is synchronized to the hospital information system, and the parameter management and auditing module records operation logs.

[0039] In the triage system's operation, the default threshold for CACS stratification is 120 points. A tertiary hospital adjusted the threshold to 150 points for elderly patients (≥65 years old). The system stably adapted and generated corresponding stratification results, achieving both flexibility and case accuracy. The algorithm performance verification process involved selecting 200 independent test samples (not involved in model training). CACS scores were calculated using both the system and manual measurement. The results showed a Pearson correlation coefficient of 0.978, indicating good consistency. In clinical application, a 3-month pilot program was conducted at a chest pain center, including 500 suspected coronary artery disease patients. 320 patients (64%) in the low-risk stratum underwent priority CTA examination, of which only 82 (25.6%) underwent CAG due to positive CTA. 180 patients (36%) in the high-risk stratum underwent CAG directly, of which 126 (70%) were diagnosed requiring interventional treatment. Compared to before the pilot program, the invasive CAG examination rate decreased by 28%, the average emergency triage time was shortened by 40 minutes, and the utilization rate of medical resources increased by 35%.

[0040] The aforementioned functional modules have corresponding extended functions, specifically including a data access module that supports multimodal data expansion, such as the ability to access electrocardiograms and blood biochemical indicators to enhance risk stratification dimensions; an AI score calculation module that supports algorithm model iteration and updates, and can access new calcification scoring standards, such as Volume scores and Mass scores; and a rules engine module that can import the latest clinical guidelines (such as ESC and ACC / AHA guidelines) and automatically update stratification rules to improve system timeliness.

[0041] This technological solution can be widely applied in clinical departments such as chest pain centers, cardiology departments, and radiology departments in hospitals at all levels. It is particularly suitable for scenarios such as rapid triage of suspected coronary heart disease patients in the emergency department, coronary heart disease risk screening in physical examination centers, and guidance for coronary heart disease diagnosis in primary hospitals. The system requires no modification to existing medical equipment, has low deployment costs, and is easy to operate. It can effectively improve the efficiency and accuracy of diagnosis and treatment in medical institutions, reduce medical risks and resource waste, and has broad market prospects and significant socio-economic benefits.

[0042] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A triage system for coronary heart disease examination pathways based on AI technology, characterized in that: The system includes a data access module, an AI score calculation module, a symptom assessment module, a rule engine module, a result display and interface module, and a parameter management and auditing module. The system performs the following steps: S1. Collect non-ECG-gated chest CT image data of suspected coronary heart disease patients through the data access module and transmit them to the AI ​​analysis platform in DICOM standard format; S2. The AI ​​scoring module uses a deep learning segmentation model to first segment the coronary artery region and then filter calcified regions according to a preset threshold. It automatically identifies the anatomical region of the coronary artery and calcified lesions, calculates and outputs the total CACS score and the scores of the optional left main coronary artery (LM), left anterior descending artery (LAD), circumflex artery (LCX), and right coronary artery (RCA) branches. S3. Based on three preset core features, the symptom judgment module determines whether the patient's chest pain is typical or atypical. S4. Based on AI-CACS scores, chest pain types, and evidence of the compatibility between CACS and examination methods, the rule engine module constructs a binary risk stratification system and outputs low-risk or high-risk results. S5. Based on the risk stratification results, corresponding treatment pathway suggestions are generated. For low-risk cases, coronary CTA is recommended as the priority, while for high-risk cases, direct invasive coronary angiography (CAG) is recommended. S6. Visualize the analysis results through the results display and interface module, and write key data into the hospital's information system.

2. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The three core characteristics in step S3 are: retrosternal discomfort with nature and duration consistent with typical coronary heart disease pain; pain can be induced by exertion or emotional stress; pain is relieved within minutes after rest and / or nitrate treatment; all three characteristics are consistent for typical chest pain, while only two or fewer characteristics are consistent for atypical chest pain.

3. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The binary risk stratification rule in step S4 is as follows: the low-risk stratification is defined as 0 ≤ CACS ≤ 120 points and atypical chest pain; the high-risk stratification is defined as CACS > 120 points or typical chest pain. The CACS stratification threshold supports personalized configuration by medical institutions.

4. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The deep learning segmentation model of the AI ​​integral calculation module in step S2 is a convolutional neural network (CNN) model with a preset density threshold of ≥130HU. The Agatston scoring algorithm is used to calculate the CACS score. The Pearson correlation coefficient between this model and the ECG-gated calcium integral scan results is ≥0.985, and the intragroup correlation coefficient (ICC) for risk stratification is ≥0.

942.

5. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: When the CTA examination in step S5 indicates vascular stenosis ≥50%, the system automatically triggers a subsequent CAG examination recommendation; when the CAG examination in step S5 indicates severe stenosis, the system provides information on the indications for interventional treatment.

6. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The data access module in step S1 adopts the DICOM standard interface, supports data access from multiple brands of CT equipment, and has functions for patient queue classification, diagnosis and treatment task creation, and status tracking.

7. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The symptom assessment module provides a structured chest pain feature input interface, supporting manual input or data extraction in conjunction with an electronic medical record (EMR) system.

8. The coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The result display and interface module in step S6 automatically writes key data into the hospital's HIS / LIS / PACS system through the HL7 interface. The visualized report includes CACS score, chest pain type, risk stratification results, examination recommendations, and evidence.

9. A coronary heart disease examination pathway triage system based on AI technology according to claim 1, characterized in that: The parameter management and auditing module provides a visual parameter configuration interface, supports adjusting CACS stratification thresholds and CTA positive thresholds, and records model version iteration logs, user operation trajectories, and data access records to meet medical data security compliance requirements.

10. A coronary heart disease examination pathway triage system based on AI technology according to any one of claims 1-9, characterized in that: The system supports in-hospital or cloud deployment, and can be expanded to access multimodal data such as electrocardiograms and blood biochemical indicators. It also supports iterative updates of algorithm models and import of clinical guidelines.

Images