System for assessing hemodynamic significance of coronary stenosis

Abstract

The application relates to the field of biological diagnosis and discloses a coronary artery stenosis hemodynamic significance evaluation system, which comprises an image acquisition module, an image processing module, an image segmentation and parameter extraction module, a hemodynamic evaluation module and the like. The image segmentation and parameter extraction module is used for extracting CCTA anatomical parameters, the parameters comprising a Duke risk score and a minimum lumen diameter. The hemodynamic evaluation module is used for calculating the ratio of the Duke risk score and the minimum lumen diameter (DJS / MLD), comparing the ratio with a preset threshold, outputting a result that the coronary artery stenosis is a high-probability event without hemodynamic significance when DJS / MLD<=1.96, and outputting a result that the coronary artery stenosis is a high-probability event with hemodynamic significance when DJS / MLD>1.96. The application solves the technical problems that existing invasive FFR examination has high risk and high cost, and traditional CCTA only provides anatomical information and cannot evaluate functional significance.

Description

Technical Field

[0001] This application relates to the field of biological diagnostics, and more specifically, to a system for assessing the hemodynamic significance of coronary artery stenosis. Background Technology

[0002] Coronary artery disease (CAD) is a heart condition primarily caused by narrowing of the coronary arteries due to factors such as atherosclerosis, leading to insufficient blood supply to the myocardium. According to guidelines from the American College of Cardiology / American Heart Association (ACC / AHA) and the European Society of Cardiology (ESC), coronary CT angiography (CCTA) has been established as a first-line diagnostic method for CAD due to its non-invasiveness and excellent negative predictive value. CCTA not only visualizes the coronary artery lumen but also clearly shows atherosclerotic plaques, providing information on plaque composition, both of which are independently associated with adverse clinical outcomes. However, CCTA only provides anatomical information and cannot accurately reflect hemodynamic conditions.

[0003] Fractional flow reserve (FFR) is the gold standard for invasively evaluating the hemodynamic significance of coronary artery lesions. Compared to FFR, the degree of stenosis assessed by either CCTA or coronary angiography is partially mismatched with the hemodynamic significance of the lesion. Multiple prospective, multicenter clinical studies have demonstrated that FFR-guided treatment strategies, compared to those based solely on angiographic stenosis, better improve patient outcomes, prolong event-free survival, and reduce unnecessary percutaneous coronary interventions (PCIs). Therefore, current clinical assessment of obstructive coronary arteries (especially borderline lesions) relies more on invasive or non-invasive functional examinations (FFR, myocardial perfusion imaging, etc.) rather than solely on the degree of stenosis shown on angiography. FFR (Fiber Retention Rate) is a technique for assessing the severity of coronary artery stenosis from a physiological perspective. During transcatheter coronary angiography, it measures the pressure distal to the stenotic lesion (guidewire pressure) and proximal to the stenotic lesion (aortic pressure) at maximum blood flow (congestion), and calculates the pressure ratio. The result is the proportion of blood flow through the lesion, i.e., the FFR value, with a normal value of 1. Currently, an FFR of ≤0.8 is considered to indicate that the lesion may cause myocardial ischemia. However, FFR measurement using a pressure guidewire is invasive, expensive, and may cause adverse reactions in a small number of patients. Furthermore, regular invasive follow-up examinations for patients with coronary artery disease are not clinically feasible.

[0004] In view of this, it is indeed necessary to provide a new non-invasive hemodynamic assessment system to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to provide a non-invasive and highly accurate CCTA-based system for assessing the hemodynamic significance of coronary artery stenosis, in order to solve the technical problems of high risk and high cost of existing invasive FFR examinations, and the fact that traditional CCTA only provides anatomical information and cannot assess functional significance.

[0006] To achieve the above-mentioned objectives, this application provides the following technical solution:

[0007] In a first aspect, this application provides a system for assessing the hemodynamic significance of coronary artery stenosis, comprising: The image acquisition module is used to acquire raw images from a CCTA examination using a dual-source, dual-width CT system. The image processing module is used to reconstruct high-resolution CCTA images using the ePhase algorithm, CardioCapture motion correction, and CardioBoost deep learning noise reduction algorithm. The image segmentation and parameter extraction module is used to segment CCTA images, obtain target segmented images, and extract CCTA anatomical parameters, including Duke risk score and minimum lumen diameter. The hemodynamic assessment module calculates the Duke risk score to minimum lumen diameter ratio (DJS / MLD) and compares this ratio with a preset threshold. When DJS / MLD ≤ 1.96, the output result indicates that the coronary artery stenosis is a high-probability event without hemodynamic significance; when DJS / MLD > 1.96, the output result indicates that the coronary artery stenosis is a high-probability event with hemodynamic significance.

[0008] Furthermore, the raw CCTA images acquired by the image acquisition module were obtained by the United Imaging Healthcare dual-width coverage CT system uCT SiriuX Elite.

[0009] Furthermore, the CCTA anatomical parameters extracted by the image segmentation and parameter extraction module also include at least one of the following: lesion length, minimum lumen area, diameter stenosis rate, area stenosis rate, plaque burden, remodeling index, and Duke risk score.

[0010] Furthermore, the Duke hazard score is calculated using the following method: The Duke Jeopardy score was used to assess the extent of perfused myocardial tissue innervated by the target stenosis. The coronary tree was divided into 6 segments: left anterior descending artery, the largest diagonal branch, the largest septal perforator, left circumflex artery, the largest obtuse marginal branch, and posterior descending artery. Each segment was assigned 2 points. For the left anterior descending artery, if the lesion is located in a well-developed area, i.e. before the largest diagonal branch with a diameter ≥2mm, or within the proximal 1 / 3 of the vessel, it is considered a proximal lesion; the rest are considered distal lesions. If it is a left-dominant type, the right coronary artery is not scored, and the left circumflex artery is scored with an additional 2 points. All segments located distal to the index stenosis are considered "danger zones", and the highest possible score for the entire myocardium is 12 points. (1) When the lesion is located in the proximal part of the left anterior descending artery, before the thickest septal branch and the thickest diagonal branch, the score is 6 points; (2) When the lesion is located after the largest septal branch and before the largest diagonal branch, the score is 4 points; (3) Right coronary dominant type: When the lesion is located in the proximal segment of the left circumflex artery and before the thickest obtuse marginal branch, the score is 4 points; (4) Right coronary dominant type: When the lesion is located in the proximal or middle segment of the right coronary artery, before the posterior descending artery originates, the score is 2 points; (5) If it is left dominant, the right coronary artery is not scored, and the left circumflex artery is scored by 2 points.

[0011] Secondly, this application provides the application of the coronary artery stenosis hemodynamic significance assessment system in the preparation of products for predicting whether coronary artery stenosis causes myocardial ischemia.

[0012] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the functions of the system.

[0013] In summary, this application has the following beneficial effects: Functional assessment of the hemodynamic significance of coronary artery stenosis is crucial for guiding treatment decisions in coronary artery disease (CAD). Currently, fractional flow reserve (FFR) is the gold standard for invasively evaluating the hemodynamic significance of coronary artery lesions. Compared to angiography-guided percutaneous coronary intervention (PCI) alone, FFR-guided PCI has been shown to improve clinical outcomes over a five-year follow-up period. FFR is a technique that assesses the severity of coronary artery stenosis from a physiological perspective. During transcatheter coronary angiography, it measures the pressure distal to the stenotic lesion (guidewire pressure) and the pressure proximal to the stenotic lesion (aortic pressure) at maximum flow (congestion), and calculates the pressure ratio to determine the proportion of blood flow through the lesion. However, FFR measurement using a pressure guidewire is invasive, expensive, and may cause adverse reactions in a small number of patients. Furthermore, regular invasive follow-up examinations for CAD patients are not clinically feasible.

[0014] This application proposes for the first time a novel, non-invasive, and highly accurate hemodynamic assessment system for obtaining coronary plaque location based on dual-source, dual-width CCTA. CardioBoost deep learning denoising utilizes a deep learning network trained to extract semantic information from images, thereby reconstructing high-quality images, which is particularly crucial in low-dose scanning. Finally, the trained CardioBoost deep learning network can be easily integrated into routine clinical workflows. In each training cycle, all simulated low-dose images are randomly shuffled and input into the network. A dedicated loss function quantitatively measures the pixel-level difference between the network output and the real, clear image. This error metric drives the optimization process through backpropagation and adaptive gradient descent, systematically adjusting network parameters to minimize reconstruction error. Millions of parameters in the deep learning network are tuned through backpropagation to reduce the difference between the network output and normal-dose images. Subsequently, the algorithm is validated using a large amount of real low-dose dataset not used for training to ensure its robustness and accuracy. Under the same scanning conditions, CardioBoost offers higher spatial resolution compared to traditional filtered backprojection, significantly reducing image noise and enhancing image contrast. This results in clearer visualization of small vessels and calcified plaques, thus improving diagnostic confidence. The "DJS / MLD ratio" obtained from dual-source, dual-width CCTA, with its excellent diagnostic accuracy and specificity, can precisely exclude hemodynamically significant lesions, thereby reducing unnecessary invasive imaging examinations for patients and acting as a "gatekeeper." Furthermore, this index can serve as an effective alternative when FFR cannot be measured or is contraindicated. The advantage of the "DJS / MLD ratio" lies in combining anatomical and functional imaging with CCTA, assessing lesions from an anatomical-physiological perspective without the need for drug-induced imaging, avoiding adverse patient reactions, and eliminating the need for additional scans or increased radiation doses. This enables precise diagnosis of coronary artery disease, providing accurate anatomical and functional information for subsequent treatment planning. Attached Figure Description

[0015] Figure 1 Schematic diagram of coronary artery tree decomposition (posterior descending branch); Figure 2 Schematic diagram of the coronary artery tree (left anterior descending artery, diagonal branches, septal perforators); Figure 3 Schematic diagram of the coronary artery tree (left circumflex branch, obtuse marginal branch); Figure 4Example images; (A) Maximum density projection image shows a focal severe stenosis at the opening of the obtuse marginal branch (white arrow), with a Duke risk score of 2; the small image in the upper right corner indicates that the minimum lumen diameter was manually measured to be 1.4 mm; Duke risk score / minimum lumen diameter = 2 / 1.4 = 1.43, which is <1.96, suggesting that the stenosis is unlikely to have hemodynamic significance; (B) Invasive coronary angiography shows severe stenosis at the opening of the obtuse marginal branch, and the gold standard fractional flow reserve (FFR) measured by pressure guidewire is 0.91, confirming that the lesion does not have functional stenosis. Detailed Implementation

[0016] The technical solutions and effects of this application will be further described in detail below with reference to the embodiments and accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining the invention and are not intended to limit the invention.

[0017] Functional assessment of the hemodynamic significance of coronary artery stenosis is crucial for guiding treatment decisions in coronary artery disease. Coronary hemodynamics is determined by multiple factors, including the degree of stenosis, myocardial mass, and microvascular resistance, especially the extent of downstream myocardial tissue supplied by the stenotic lesion. For the same degree of coronary artery stenosis, a larger downstream myocardial perfusion area is more likely to cause myocardial ischemia. Therefore, adding parameters related to myocardial perfusion area to the anatomical stenosis can reduce the mismatch between anatomy and hemodynamics. Calculating the extent of myocardial tissue requires specialized post-processing software and is a cumbersome process, requiring delineation of the endocardium and epicardium to achieve accurate extraction of myocardial tissue. The Duke Jeopardy score (DJS) is a simple and rapid method for calculating myocardial tissue extent, initially used in invasive imaging to calculate the extent of downstream myocardial tissue supplied by the stenotic lesion. However, the fractional flow reserve of the coronary artery is affected by both the volume of myocardial blood supply distal to the stenosis and the minimum luminal area at the lesion site. Therefore, we propose for the first time to calculate functional coronary artery stenosis using the DJS / MLD ratio obtained from CCTA. This ratio simultaneously incorporates the extent of downstream myocardial tissue innervated by the stenotic site and the minimum luminal area at the lesion. Our results show that DJS / MLD is superior to other traditional anatomical parameters in identifying functional coronary artery stenosis. Adding the parameter of myocardial perfusion range (DJS) to the anatomical stenosis (MLD) reduces the mismatch between anatomy and hemodynamics. This application is the first to use the novel parameter "DJS / MLD ratio" to assess the hemodynamics of coronary artery stenosis.

[0018] Example This study included patients clinically suspected of having coronary artery disease who underwent CCTA and FFR examinations during the period from January 2018 to December 2019. FFR measurement is used to determine the hemodynamic significance of coronary artery stenosis and guide treatment strategies. If the FFR is less than or equal to 0.8, it indicates that the lesion at the stenosis site has hemodynamic significance and causes myocardial ischemia, and revascularization is recommended. If the FFR is greater than 0.8, it indicates that the coronary artery lesion has no hemodynamic significance, and conservative drug treatment is recommended.

[0019] Inclusion requirements: Subjects must consent to undergo CCTA examination before undergoing invasive FFR examination.

[0020] Exclusion criteria: 1) Previous coronary artery bypass grafting or any coronary intervention; 2) Multiple lesions in the target vessel; 3) History of myocardial infarction; 4) Poor CCTA image quality, unable to be evaluated; 5) Diffuse calcification of the target lesion (defined as calcification curvature greater than 180 degrees in any cross-section of the lesion); 6) ICA showing complete coronary artery occlusion or retrograde collateral flow; 7) Comorbid valvular heart disease, cardiomyopathy, or NYHA class IV heart failure; 8) Intracardiac metallic devices; 9) Congenital coronary artery malformations.

[0021] All eligible participants underwent CCTA scanning, as follows: All patients underwent CCTA using a dual-wide coverage CT system (uCT SiriuX Elite, United Imaging Healthcare, Shanghai, China) with a prospective ECG-gated axial acquisition protocol. The scanner was equipped with two X-ray tubes (tube A and tube B) and two corresponding detector arrays, operating with identical parameters during the scan. Both tube voltages were set to 100 kVp with automatic tube current modulation (reference tube current 150 mAs). The gantry rotation time was 0.229 seconds, and the acquisition window covered 30%–80% of the cardiac cycle to encompass both systole and diastole. The scan range covered the entire cardiac region from the carina to the diaphragm, with Z-axis coverage of 12, 14, or 16 cm depending on the patient's heart size. The field of view (FOV) was 420 mm × 420 mm, and the matrix size was 512 × 512. All participants were pre-treated with medication based on baseline heart rate: patients with a heart rate (HR) greater than 75 bpm received an oral beta-blocker 10 minutes prior to the scan. Simultaneously, all subjects took nitroglycerin sublingually 5 minutes before the start of the CCTA scan. Contrast agent (iodine concentration = 350 mgI / mL) was injected intravenously at a dose calculated as body weight (kg) × 0.7 mL / kg, with an injection flow rate of 5 mL / s, followed by 30 mL of normal saline at a flow rate of 5 mL / s. Image acquisition was initiated using bolus tracking technology, with the region of interest placed within the descending aorta; scanning was automatically triggered when the CT value reached 120 HU. Images were reconstructed with a slice thickness of 0.33 mm and a reconstruction interval. Automatic cardiac cycle phase selection was performed using an AI-based ePhase algorithm, which identifies the phase with the least coronary motion artifacts within the entire cardiac cycle or the diastolic / systolic window. AI-based motion correction technology (CardioCapture, United Imaging Healthcare) was applied to reduce coronary motion artifacts. All images were reconstructed using a deep learning denoising algorithm (CardioBoost).

[0022] CardioBoost deep learning-based denoising utilizes a trained deep learning network capable of extracting semantic information from images to reconstruct high-quality images, a crucial feature, especially in low-dose scanning. Finally, the trained CardioBoost deep learning network can be easily integrated into routine clinical workflows.

[0023] In each training epoch, all simulated low-dose images are randomly shuffled and fed into the network. A dedicated loss function quantitatively measures the pixel-level difference between the network output and the real, sharp images. This error metric drives the optimization process through backpropagation and adaptive gradient descent, systematically tuning the network parameters to minimize reconstruction error. Millions of parameters in the deep learning network are tuned via backpropagation to reduce the difference between the network output and normal-dose images. Subsequently, the algorithm is validated using a large dataset of real low-dose images not used for training to ensure its robustness and accuracy.

[0024] Under the same scanning conditions, CardioBoost has higher spatial resolution than traditional filtered backprojection, which can significantly reduce image noise and enhance image contrast. Therefore, it can display small blood vessels and calcified plaques more clearly, thus improving diagnostic confidence.

[0025] Image processing Data was transferred to an offline workstation, and cardiac phases with the best image quality were selected for subsequent analysis. CCTA images were obtained; image segmentation was performed on the CCTA images to obtain target segmented images; wherein, the target segmented images contain multiple coronary artery anatomical structures. These included the right coronary artery, posterior descending artery, posterior left ventricular branch, left main coronary artery, left anterior descending artery, left circumflex artery, diagonal branch, and obtuse marginal branch.

[0026] Parameters obtained by CCTA include: lesion length, minimum lumen diameter (MLD), minimum lumen area (MLA), diameter stenosis (DS), area stenosis (AS), plaque burden, remodeling index, Duke Jeopardy score (DJS), and Duke Jeopardy score / Minimal lumendiameter (DJS / MLD) ratio.

[0027] The parameters are explained in detail below: Lesion length (LL): The length of a lesion is the distance from its proximal end to its distal end. Minimum lumen diameter (MLD): The minimum lumen diameter is manually measured at the narrowest point of the lesion on a cross-sectional image. Minimum lumen area (MLA): The minimum lumen area is manually measured at the narrowest point of the lesion on a cross-sectional image. Diameter stenosis (DS): The ratio of the smallest diameter of the vessel lumen in the lesion segment to the average diameter of the lumen of the reference vessels proximal and distal to the lesion. Area stenosis (AS): (mean area of ​​the lumen of the reference vessels proximal and distal to the lesion - minimum lumen area) / mean area of ​​the lumen of the reference vessels proximal and distal to the lesion; Plaque burden: (Cross-sectional area of ​​the vessel at the narrowest point of the lesion - Minimum lumen area) / Cross-sectional area of ​​the vessel at the narrowest point of the lesion; Remodeling index: the maximum lumen diameter at the narrowest point of the lesion and the average lumen diameter of the reference vessels proximal and distal to the lesion; The Duke Jeopardy score (DJS) assesses the extent of perfused myocardial tissue innervated by the target stenosis. The coronary tree is divided into six segments: the left anterior descending artery (LAD), the largest diagonal branch, the largest septal perforator, the left circumflex artery, the largest obtuse marginal branch, and the posterior descending artery, each assigned 2 points (Figure 1-3). For the LAD, if the lesion is located before the origin of the largest diagonal branch (≥ 2 mm in diameter) or within the proximal third of the vessel, it is considered a proximal lesion; otherwise, it is considered a distal lesion. In left-dominant cases, the right coronary artery is not scored, and the left circumflex artery receives an additional 2 points. All segments distal to the index stenosis are considered "danger zones." The maximum possible score for the entire myocardium is 12 points.

[0028] 1. When the lesion is located in the proximal part of the left anterior descending artery, before the origin of the largest septal branch and the largest diagonal branch, the score is 6 points; 2. When the lesion is located after the largest septal branch originates and before the largest diagonal branch originates, the score is 4 points; 3. Right coronary dominant type: When the lesion is located in the proximal segment of the left circumflex artery, before the largest obtuse marginal branch originates, the score is 4 points; 4. Right coronary dominant type: When the lesion is located in the proximal or middle segment of the right coronary artery, before the posterior descending artery originates, the score is 2 points; 5. If it is a left-dominant type, the right coronary artery is not scored, and the left circumflex artery is scored with an additional 2 points; Duke Jeopardy score / Minimum lumendiameter (DJS / MLD): Duke Jeopardy score divided by minimum lumen diameter.

[0029] result This prospective study included 215 patients who underwent both CCTA and invasive FFR measurement. Subsequently, 9 patients who had previously undergone revascularization, 14 patients with tandem lesions, and another 31 patients were excluded due to poor interpretation of CCTA images or diffuse calcification of the lesions.

[0030] A total of 161 patients were included [mean age: 65±10.3 (42–85) years; 109 males, mean age 67.3±10.6 (42–83) years; 52 females, mean age 61.6±9.2 (43–85) years; p= 0.073], with 175 lesions. Demographic characteristics are shown in Table 1. The mean dose-length product of CCTA was 509.7±111.7 mGy·cm (241–709 mGy·cm), and the mean effective dose was 7.1±1.5 mSv (3.5–11.4 mSv). The mean contrast agent usage for CCTA and created radiography was 82.9±9.6 mL and 159.3±32 mL, respectively. The mean radiation dose for created radiography was 54.5±23.6 Gy·cm. 2 The average operation time was 53 ± 28 min.

[0031] Table 1 Basic Clinical Data

[0032] Note: Unless otherwise stated, all data are number of patients, and percentages are in parentheses. a. Data are expressed as mean ± standard deviation; b. Data represents the number of patients, with percentages in parentheses; c represents the number of cases with lesions, and the percentage is shown in parentheses.

[0033] Based on the gold standard FFR = 0.8 as the cutoff value, lesions were divided into two groups: those with FFR ≤ 0.8 (hemodynamically significant) and those with FFR > 0.8 (hemodynamically insignificant). Based on CCTA data, lesion length, minimum luminal diameter, minimum luminal area, diameter stenosis rate, area stenosis rate, Duke risk score, and DJS / MLD ratio were all significantly higher in the hemodynamically significant group (FFR ≤ 0.8) than in the group without hemodynamic significance (FFR > 0.8) (p values ​​were all less than 0.05, see Table 2). There was no statistically significant difference in the remodeling index between the two groups (1.09 ± 0.09 vs. 1.06 ± 0.08, p = 0.057).

[0034] Table 2 Comparison of CCTA parameters between the hemodynamically significant and non-significant groups.

[0035] Note: Unless otherwise specified, data are mean ± standard deviation.

[0036] Spearman correlation analysis showed that the correlation coefficient between the minimum lumen diameter and the gold standard FFR was 0.42 (95% confidence interval: 0.29 ~ 0.53, P<0.001), the correlation coefficient between the minimum lumen area and the gold standard FFR was 0.40 (95% confidence interval: 0.29 ~ 0.53, P<0.001), the correlation coefficient between narrow diameter and the gold standard FFR was -0.38 (95% confidence interval: 0.50~ 0.24, P<0.001), the correlation coefficient between narrow area and the gold standard FFR was -0.35 (95% confidence interval: 0.48 ~ The correlation coefficient between plaque load and the gold standard FFR was -0.30 (95% confidence interval: 0.22, P<0.001). 0.43 ~ 0.16, P<0.001), the correlation coefficient between lesion length and the gold standard FFR was -0.30 (95% confidence interval: 0.43 ~ 0.16, P<0.001), based on the correlation coefficient between the degree of stenosis in the creative imaging and the gold standard FFR, which is -0.47 (95% confidence interval: 0.58 ~ The correlation coefficient between the Duke risk score and the gold standard FFR was -0.35 (95% confidence interval: 0.35, P<0.001). 0.46 ~ The correlation coefficient between the Duke risk score / minimum lumen diameter ratio and the gold standard FFR was -0.57 (95% confidence interval: 0.21, P<0.001). 0.66 ~ 0.46, P<0.001), among which the Duke risk score / minimum lumen diameter ratio had the strongest correlation with the gold standard FFR correlation coefficient.

[0037] Among all enrolled individuals, the area under the curve (AUC) for the DJS / MLD ratio was the largest (AUC = 0.863, p < 0.001). Compared with other parameters, the DJS / MLD ratio had the highest diagnostic accuracy, specificity, positive predictive value, and negative predictive value in predicting the significance of coronary artery hemodynamics, with an optimal cutoff value of 1.96 (Table 4). ROC curve analysis showed that, regardless of whether it was the right or left coronary artery, the DJS / MLD ratio was superior to all other parameters in predicting the significance of coronary artery stenosis hemodynamics (Table 3). The AUCs for DJS / MLD in the right and left coronary arteries were 0.822 and 0.851, respectively, with no statistically significant difference (p = 0.713).

[0038] Five-fold cross-validation was used to validate the optimal critical value of 1.96 for DJS / MLD; the mean AUC of DJS / MLD obtained from five-fold cross-validation was 0.861 (95% CI: 0.817–0.904). Based on the observations used for modeling, the mean AUC of the proposed model for DJS / MLD was 0.863 (95% CI: 0.852–0.874).

[0039] Table 3. Area under the receiver operating characteristic curve for diagnosing significant and non-significant hemodynamic stenosis.

[0040] This specific embodiment is merely an explanation of this application and is not intended to limit it. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they fall within the scope of the claims of this application.

Claims

1. A system for assessing the hemodynamic significance of coronary artery stenosis, characterized in that, include: The image acquisition module is used to acquire raw images from a CCTA examination using a dual-source, dual-width CT system. The image processing module is used to reconstruct high-resolution CCTA images using the ePhase algorithm, CardioCapture motion correction, and CardioBoost deep learning noise reduction algorithm. The image segmentation and parameter extraction module is used to segment CCTA images, obtain target segmented images, and extract CCTA anatomical parameters, including Duke risk score and minimum lumen diameter. The hemodynamic assessment module calculates the Duke risk score to minimum lumen diameter ratio (DJS / MLD) and compares this ratio with a preset threshold. When DJS / MLD ≤ 1.96, the output result indicates that the coronary artery stenosis is a high-probability event without hemodynamic significance; when DJS / MLD > 1.96, the output result indicates that the coronary artery stenosis is a high-probability event with hemodynamic significance.

2. The coronary artery stenosis hemodynamic significance assessment system according to claim 1, characterized in that, The raw CCTA images acquired by the image acquisition module were obtained by the United Imaging Healthcare dual-width coverage CT system uCT SiriuX Elite.

3. The coronary artery stenosis hemodynamic significance assessment system according to claim 1, characterized in that, The CCTA anatomical parameters extracted by the image segmentation and parameter extraction module also include at least one of the following: lesion length, minimum lumen area, diameter stenosis rate, area stenosis rate, plaque burden, remodeling index, and Duke risk score.

4. The coronary artery stenosis hemodynamic significance assessment system according to claim 1, characterized in that, The Duke hazard score is calculated using the following method: The Duke Jeopardy score was used to assess the extent of perfused myocardial tissue innervated by the target stenosis. The coronary tree was divided into 6 segments: left anterior descending artery, the largest diagonal branch, the largest septal perforator, left circumflex artery, the largest obtuse marginal branch, and posterior descending artery. Each segment was assigned 2 points. For the left anterior descending artery, if the lesion is located in a well-developed area, i.e. before the largest diagonal branch with a diameter ≥2mm, or within the proximal 1 / 3 of the vessel, it is considered a proximal lesion; the rest are considered distal lesions. If it is a left-dominant type, the right coronary artery is not scored, and the left circumflex artery is scored with an additional 2 points. All segments located distal to the index stenosis are considered "danger zones", and the highest possible score for the entire myocardium is 12 points. (1) When the lesion is located in the proximal part of the left anterior descending artery, before the thickest septal branch and the thickest diagonal branch, the score is 6 points; (2) When the lesion is located after the largest septal branch and before the largest diagonal branch, the score is 4 points; (3) Right coronary dominant type: When the lesion is located in the proximal segment of the left circumflex artery and before the thickest obtuse marginal branch, the score is 4 points; (4) Right coronary dominant type: When the lesion is located in the proximal or middle segment of the right coronary artery, before the posterior descending artery originates, the score is 2 points; (5) If it is left dominant, the right coronary artery is not scored, and the left circumflex artery is scored by 2 points.

5. The use of the coronary artery stenosis hemodynamic significance assessment system according to any one of claims 1-4 in the preparation of products for predicting whether coronary artery stenosis causes myocardial ischemia.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it performs the functions of the system according to any one of claims 1-4.

Images