Coronary stenosis functional evaluation method based on angiography imageomics

Abstract

The present application relates to the technical field of blood vessel image processing, and particularly relates to a coronary artery stenosis functional evaluation method based on angiography imageomics. According to the characteristic quantitative value of the stenosis degree of each patient in the multi-dimensional image performance characteristics, the stenosis grade of each patient is obtained. According to the distribution of the characteristic quantitative value of the patient in different image performance characteristics in different stenosis grades, the key possibility of the current patient in each image performance characteristic is obtained, and the image performance key characteristics are screened out. According to the distribution of the characteristic quantitative value of each image performance key characteristic of different patients in the stenosis grade where the current patient is located, the coronary artery quality contribution degree of each image performance key characteristic in the stenosis grade where the current patient is located is obtained. The coronary artery stenosis evaluation score of the stenosis grade where the current patient is located is obtained. The present application can accurately obtain the coronary artery stenosis evaluation score of the stenosis grade where the current patient is located, and improve the accuracy of the stenosis functional evaluation.

Description

Technical Field

[0001] This invention relates to the field of vascular image processing technology, and more specifically to a method for functional assessment of coronary artery stenosis based on angiography radiomics. Background Technology

[0002] Traditional methods for assessing coronary artery stenosis using angiography primarily rely on single imaging findings to determine the degree of stenosis. However, angiography cannot directly reflect the impact on myocardial blood supply function. With the development of radiomics and artificial intelligence technologies, researchers are utilizing angiographic radiomics data to extract a wealth of image features, including morphology, texture, density, and blood flow velocity. Radiomics methods can extract potential clinical information from complex image data and, combined with machine learning or deep learning algorithms, perform modeling. This allows for a more comprehensive analysis of the structural features of blood vessels and their relationship with functional ischemia, thereby providing a more accurate functional assessment.

[0003] In existing technologies, a dual-channel deep learning model is used to extract multiple related features from radiomics, and then gradient boosting trees are used to screen out key features for evaluation, output the functional ischemia probability of stenotic lesions, and complete the functional assessment of coronary artery stenosis. However, due to the high correlation between indicators such as the morphology, texture, and density of the same plaque, as well as the collinearity of morphological features, the tree model is biased towards strongly correlated features, while missing weakly correlated features that may be decisive, resulting in poor accuracy in coronary artery stenosis assessment. Summary of the Invention

[0004] To address the problem of poor accuracy in coronary artery stenosis assessment caused by the biased analysis of strongly correlated features while omitting potentially decisive features in weakly correlated features, the present invention aims to provide a functional assessment method for coronary artery stenosis based on angiography radiomics. The specific technical solution adopted is as follows: This invention proposes a method for functional assessment of coronary artery stenosis based on angiography radiomics, the method comprising: Acquire images of the patient's coronary arteries at different projection angles during end-diastole and obtain the characteristic quantification values ​​of the patient's multidimensional imaging features; Based on the feature quantification value of the stenosis degree of each patient in the multidimensional imaging features, the stenosis level of each patient is obtained; based on the distribution of feature quantification values ​​of patients in different imaging features within different stenosis levels, the key probability of the current patient in each imaging feature is obtained, and key imaging features are screened out. Based on the distribution of feature quantification values ​​of key features in each imaging manifestation of different patients in the current stenosis level, the intensity stability coefficient of each key feature in each imaging manifestation of the current stenosis level is obtained; based on the distribution of intensity stability coefficients of key features in different imaging manifestations in the current stenosis level, the coronary artery quality contribution of each key feature in each imaging manifestation of the current stenosis level is obtained. Based on the contribution of key imaging features to coronary artery quality within the current stenosis level of the patient, a coronary artery stenosis evaluation score is obtained for the current stenosis level of the patient.

[0005] Furthermore, the method for obtaining the narrowness level includes: If the characteristic quantification value of a patient's stenosis degree falls within the stenosis degree range of a certain preset stenosis level, the corresponding stenosis level of the patient will be set to the preset stenosis level.

[0006] Furthermore, the method for obtaining the key possibilities includes: The mean value of the feature quantification of each imaging feature for all patients in each stenosis level is obtained as the baseline value of the imaging feature. Based on the difference between the current patient's characteristic quantification value for each imaging feature and the baseline value of the imaging feature corresponding to the previous stenosis level, as well as the degree of fluctuation of the characteristic quantification value, the key probability of the current patient for each imaging feature is obtained.

[0007] Furthermore, the method for obtaining the key possibilities includes: Obtain the degree of fluctuation in the characteristic quantitative values ​​of each imaging feature of historical patients in each stenosis grade; The difference between the current patient's characteristic quantification value for each imaging feature and the baseline value of the imaging feature corresponding to the previous stenosis level is calculated and summed to determine the degree of fluctuation of characteristic quantification values ​​for all stenosis levels. This sum is used as the key probability for the current patient in each imaging feature.

[0008] Furthermore, the method for obtaining the key features of the image representation includes: If the probability of a key image performance feature is greater than or less than a preset fluctuation threshold, the corresponding image performance feature will be used as the key image performance feature.

[0009] Furthermore, the method for obtaining the strength stability coefficient includes: Based on the differences between the feature quantification values ​​and corresponding image feature baseline values ​​of key features in each imaging manifestation for different patients in the current stenosis level, as well as the image manifestation feature baseline values, the intensity stability coefficient of each key feature in each imaging manifestation for each stenosis level is obtained.

[0010] Furthermore, the method for obtaining the strength stability coefficient includes: The mean difference between the feature quantification value and the baseline value of the image manifestation feature for each key feature in the current stenosis level of different patients is obtained and negatively correlated, which is used as the first stability coefficient; the product of the baseline value of the image manifestation feature and the first stability coefficient is obtained as the intensity stability coefficient of each key feature in the image manifestation in each stenosis level.

[0011] Furthermore, the method for obtaining the coronary artery quality contribution includes: Based on the intensity stability coefficient distribution of key imaging features in the current stenosis level of the patient, key similarity features of each key imaging feature are obtained. Based on the differences in intensity stability coefficients between key features and different key similar features in the current stenosis level of the patient, as well as the intensity stability coefficient of each key feature in the current stenosis level, the coronary artery quality contribution of each key feature in the current stenosis level of the patient is obtained. The difference in intensity stability coefficient is positively correlated with the coronary artery quality contribution, and the intensity stability coefficient is positively correlated with the coronary artery quality contribution.

[0012] Furthermore, the method for obtaining the key similarity features includes: Obtain the relative distance of the intensity stability coefficients of different key features of image representation, select a preset number of relative distances between all key features of image representation in descending order, and use the corresponding key features of image representation as key similar features.

[0013] Furthermore, the method for obtaining the coronary artery stenosis evaluation score includes: Based on the contribution of all key imaging features to coronary artery quality in the current stenosis level of the patient, a gradient boosting tree is used for pre-training to obtain the basic contribution coefficient of each key imaging feature in the current stenosis level of the patient. The summation of the products between the contribution baseline coefficients of all key imaging features in the current stenosis grade of the patient and the contribution of coronary artery quality is obtained and normalized to serve as the coronary artery stenosis evaluation score for the current stenosis grade of the patient.

[0014] The present invention has the following beneficial effects: This invention obtains the stenosis level of each patient based on the quantified feature value of stenosis degree in multidimensional imaging features, which helps in subsequent analysis of patients with the same stenosis level. Based on the distribution of quantified feature values ​​of patients in different imaging features within different stenosis levels, it obtains the key probability of the current patient in each imaging feature, filters out key imaging features, avoids equal weighting of all features, and highlights imaging features important to the current patient. Based on the distribution of quantified feature values ​​of key features of different patients in each imaging feature within the current stenosis level, it obtains the intensity stability coefficient of each key imaging feature within the current stenosis level, quantifying the consistency of key feature changes within the stenosis level. Based on the distribution of intensity stability coefficients of key imaging features within the current stenosis level, it obtains the coronary artery quality contribution of each key imaging feature within the current stenosis level, reflecting the magnitude of the contribution of key features to coronary artery quality assessment. Based on the coronary artery quality contribution of key imaging features within the current stenosis level, it obtains the coronary artery stenosis evaluation score for the current stenosis level, more comprehensively and accurately reflecting the coronary artery stenosis characteristics of the current stenosis level. This invention improves the accuracy of functional assessment of stenosis by accurately obtaining the coronary artery stenosis evaluation score of the current stenosis level. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a method for functional assessment of coronary artery stenosis based on angiography radiomics, provided as an embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for functional assessment of coronary artery stenosis based on angiography radiomics proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] The following description, in conjunction with the accompanying drawings, details a specific scheme for a functional assessment method for coronary artery stenosis based on angiography radiomics provided by this invention.

[0020] Please see Figure 1 The diagram illustrates a flowchart of a method for functional assessment of coronary artery stenosis based on angiography radiomics, according to an embodiment of the present invention. The specific method includes: Step S1: Obtain images of the patient's coronary arteries at different projection angles during end-diastole, and obtain the feature quantification values ​​of the patient's multidimensional imaging characteristics.

[0021] In the embodiments of the present invention, considering the high correlation of indicators such as morphology, texture, and density of the same plaque, the collinearity of morphological features causes the tree model to favor strongly correlated features while omitting decisive features with weak correlations. Therefore, it is necessary to combine and analyze the structural features of blood vessels and the similarity of changes between them and the differences in the patient's stenosis level to improve the accuracy of stenosis assessment. First, information is collected using a digital subtraction angiography machine, collecting images from multiple standard projection angles to ensure that the lesion is observed from different perspectives. Considering that blood vessels regularly perform contraction and relaxation movements, the end-diastolic phase is selected for image analysis to reduce the influence of cardiac motion artifacts.

[0022] It should be noted that, in the embodiments of the present invention, in order to facilitate subsequent image processing, the known diameter of the angiography catheter is used as a scale to calibrate the pixel spacing of the image, ensuring that the subsequently measured blood vessel size, lesion length, etc. are real-world physical dimensions; anisotropic diffusion filtering is used to smooth image noise while preserving the edge information of blood vessels as much as possible; the specific means are well known to those skilled in the art and will not be described in detail here.

[0023] It should be noted that the method for obtaining the feature quantification values ​​of the patient's multidimensional imaging performance includes: analyzing the morphological and grayscale features of the coronary arteries in images from all projection angles; segmenting the coronary artery regions in the patient's images using a pre-trained U-Net neural network; extracting vascular morphological features using a geometric measurement algorithm, which includes at least the degree of stenosis and plaque volume. The degree of stenosis is calculated as the average ratio between the difference between the preset reference inner diameter of the vessel and the inner diameter of the vessel region at all projection angles and the preset reference inner diameter; and utilizing images from all projection angles... A three-dimensional mesh model of the blood vessel is reconstructed, plaque segments are identified on the 3D model, and their volumes are calculated. Radiomics features are extracted using texture analysis and wavelet transform algorithms, including at least gray-level co-occurrence matrix features and wavelet texture features. The vascular morphological features are combined with the radiomics features to form multi-dimensional image performance features. To avoid the influence of the dimensions of subsequent image performance features, the obtained results need to be Z-score standardized to obtain feature quantification values. The preset reference vascular diameter is the vascular diameter under normal conditions obtained based on relevant clinical knowledge. The specific methods are well known to those skilled in the art and will not be elaborated here.

[0024] Step S2: Based on the feature quantification value of the stenosis degree of each patient in the multidimensional imaging features, obtain the stenosis level of each patient; based on the distribution of feature quantification values ​​of patients in different imaging features within different stenosis levels, obtain the key probability of the current patient in each imaging feature, and screen out the key imaging features.

[0025] The characteristic quantitative value of the degree of stenosis reflects the functional characteristics of coronary artery stenosis. The greater the degree of stenosis, the more severe the stenosis. Thus, the stenosis grade of the patient is quantified. The stenosis grade of each patient is obtained based on the characteristic quantitative value of the degree of stenosis in the multidimensional imaging features.

[0026] Preferably, in one embodiment of the present invention, the method for obtaining the narrowness level includes: If the characteristic quantification value of a patient's stenosis degree falls within the stenosis degree range of a certain preset stenosis level, the corresponding stenosis level of the patient will be set to the preset stenosis level.

[0027] It should be noted that the greater the degree of narrowing, the more frequent the narrowing occurs, and the higher the level of narrowing. In the embodiments of the present invention, four preset narrowing levels are set based on relevant historical experience, including mild narrowing, moderate narrowing, moderate to severe narrowing, and severe narrowing, which correspond to the narrowing range of 0-50%, 50%-70%, 70%-90%, and 90%-100% in sequence. In other embodiments of the present invention, the preset narrowing levels can be set according to specific circumstances, which will not be limited or elaborated here.

[0028] Feature quantification values ​​reflect the performance status of each dimension of image manifestation features in relation to stenotic function. The more significant the change in feature quantification values ​​between each grade and adjacent grades, the more sensitive the fluctuation of image manifestation features within the stenosis grade, and the more important the feature is. Based on the distribution of feature quantification values ​​of different image manifestation features in patients within different stenosis grades, the key probability of the current patient in each image manifestation feature is obtained, and key image manifestation features are screened out.

[0029] Preferably, in one embodiment of the present invention, the method for obtaining key possibilities includes: The mean value of the feature quantification of each imaging feature for all patients in each stenosis level is obtained as the baseline value of the imaging feature. Based on the difference between the current patient's characteristic quantification value for each imaging feature and the baseline value of the imaging feature corresponding to the previous stenosis level, as well as the degree of fluctuation of the characteristic quantification value, the key probability of the current patient for each imaging feature is obtained.

[0030] Preferably, in one embodiment of the present invention, the method for obtaining key possibilities includes: Obtain the degree of fluctuation in the characteristic quantitative values ​​of each imaging feature of historical patients in each stenosis grade; The difference between the current patient's characteristic quantification value for each imaging feature and the baseline value of the imaging feature corresponding to the previous stenosis level is calculated and summed up as the sum of the fluctuations in the characteristic quantification values ​​for all stenosis levels. This sum is used as the key probability for the current patient in each imaging feature.

[0031] It should be noted that, in the embodiments of the present invention, in order to avoid the fluctuation level being 0 and the formula being meaningless, a preset adjustment threshold of a very small positive number needs to be added to the denominator. The value of the threshold can be set according to the range of the denominator, and will not be limited or elaborated here.

[0032] It should be noted that, in one embodiment of the present invention, the degree of fluctuation is characterized by calculating the standard deviation. The larger the standard deviation, the greater the degree of fluctuation, and the smaller the standard deviation, the smaller the degree of fluctuation. In other embodiments of the present invention, the degree of fluctuation can also be characterized by calculating the variance. The specific means are well known to those skilled in the art and will not be described in detail here.

[0033] The greater the likelihood of a key feature, the more significant the changes in its image characteristics compared to others, thus allowing for the selection of key image characteristics.

[0034] Preferably, in one embodiment of the present invention, the method for obtaining key features of image performance includes: if the key probability of an image performance feature is greater than or less than a preset fluctuation threshold, the corresponding image performance feature is taken as a key feature of image performance.

[0035] It should be noted that, in one embodiment of the present invention, the preset fluctuation threshold is set to 0.75; in other embodiments of the present invention, the preset fluctuation threshold may be set according to specific circumstances, and will not be limited or elaborated here.

[0036] Step S3: Based on the distribution of feature quantification values ​​of key features in each imaging manifestation of different patients in the current stenosis level, obtain the intensity stability coefficient of key features in each imaging manifestation of the current stenosis level; based on the distribution of intensity stability coefficients of key features in different imaging manifestations in the current stenosis level, obtain the coronary artery quality contribution of key features in each imaging manifestation of the current stenosis level.

[0037] For patients of the same stenosis level, the more concentrated the imaging manifestation status of the same key imaging feature, that is, the more uniform the distribution of feature quantification values, the more consistent the intensity imaging manifestation, and the greater the intensity stability coefficient; therefore, based on the feature quantification value distribution of different patients in each key imaging feature of the current stenosis level, the intensity stability coefficient of each key imaging feature in the current stenosis level can be obtained.

[0038] Preferably, in one embodiment of the present invention, the method for obtaining the strength stability coefficient includes: Based on the differences between the feature quantification values ​​and the corresponding image performance feature baseline values ​​of different patients in each image performance key feature at the current stenosis level, as well as the image performance feature baseline values, the intensity stability coefficient of each image performance key feature in each stenosis level is obtained. The difference between the feature quantification value and the image performance feature baseline value is negatively correlated with the intensity stability coefficient, while the image performance feature baseline value is positively correlated with the intensity stability coefficient.

[0039] It should be noted that the difference between the feature quantification value and the image manifestation feature baseline value represents the absolute value of the difference between the feature quantification value and the image manifestation feature baseline value. This reflects the degree of deviation of each patient's feature quantification value for each key image manifestation feature relative to the overall patient mean. The greater the difference, the more uneven the distribution of feature quantification values ​​for each key image manifestation feature among different patients, and the more unstable the intensity of the key image manifestation feature. Conversely, the larger the image manifestation feature baseline value, the greater the intensity of the same key image manifestation feature across multiple patients, the smaller the degree of variation in image manifestation among patients, and the larger the intensity stability coefficient. Therefore, the difference between the feature quantification value and the image manifestation feature baseline value is negatively correlated with the intensity stability coefficient, while the image manifestation feature baseline value is positively correlated with the intensity stability coefficient.

[0040] In one embodiment of the present invention, the mean difference between the feature quantification value and the baseline value of the image performance feature of different patients in each key image feature at the current stenosis level is obtained, and a negative correlation mapping is performed as a first stability coefficient; the product of the baseline value of the image performance feature and the first stability coefficient is obtained as the intensity stability coefficient of each key image feature in each stenosis level; for the current stenosis level... The formula for the strength stability coefficient is expressed as: ;in, Indicates the first The first of the narrow levels Intensity stability coefficient of key features in image representation; Indicates the first All patients in each of the stenosis grades were in the first... The average value of the quantified feature of each key feature of an image, i.e., the baseline value of the image performance feature; Indicates the first The first of the narrow levels The patient in The quantified value of the key features of an image; Indicates the first The number of patients within each stenotic grade; Represents an exponential function with the natural constant as its base; This indicates that the absolute value is being calculated.

[0041] It should be noted that, in one embodiment of the present invention, an exponential function with a base of the natural constant is used. The smaller the mean difference, the larger the stability coefficient after negative correlation mapping. In other embodiments of the present invention, negative correlation mapping can also be performed by taking the reciprocal. In order to avoid the denominator of the formula being 0 and the formula being meaningless, a non-zero adjustment threshold is artificially added to the denominator. The value of the threshold is specifically set according to the range of the denominator. The specific means are well known to those skilled in the art and will not be described in detail here.

[0042] The intensity stability coefficient reflects the significance and stability of key imaging features. The larger the intensity stability coefficient, the higher and more stable the quality among multiple key imaging features, and the greater the contribution of the intensity stability coefficient to the degree of coronary artery stenosis within the stenosis grade. Based on the distribution of intensity stability coefficients of key imaging features in different stenosis grades of the current patient, the coronary artery quality contribution of each key imaging feature in the current stenosis grade of the current patient can be obtained.

[0043] Preferably, in one embodiment of the present invention, the method for obtaining the coronary artery quality contribution includes: Based on the intensity stability coefficient distribution of key imaging features in the current stenosis level of the patient, obtain the current similar key features for each key imaging feature; Preferably, in one embodiment of the present invention, the method for obtaining key similar features includes: Obtain the relative distance of the intensity stability coefficients of different key features of image representation, select a preset number of relative distances between all key features of image representation in descending order, and use the corresponding key features of image representation as key similar features.

[0044] It should be noted that in the embodiments of the present invention, the relative distance is obtained by existing distance calculation methods such as Euclidean distance or Manhattan distance. The specific means are well known to those skilled in the art and will not be described in detail here.

[0045] It should be noted that in one embodiment of the present invention, the preset quantity is 3; in other embodiments of the present invention, the preset quantity can be set according to specific circumstances, and will not be limited or elaborated here.

[0046] Based on the differences in intensity stability coefficients between key features and different key similar features in the current stenosis level of the patient, as well as the intensity stability coefficient of each key feature in the current stenosis level, the coronary artery quality contribution of each key feature in the current stenosis level of the patient is obtained. The difference in intensity stability coefficient is positively correlated with the coronary artery quality contribution, and the intensity stability coefficient is positively correlated with the coronary artery quality contribution.

[0047] It should be noted that a larger intensity stability coefficient indicates more similarities in the development stages of key imaging features, greater feature significance and stability, and a greater contribution to coronary artery quality. Conversely, a smaller intensity stability coefficient indicates higher sensitivity to environmental changes and a smaller contribution to coronary artery quality. The difference in intensity stability coefficients represents the absolute value of the difference between them, reflecting the intensity stability of each key imaging feature relative to all other key similar features. A larger difference indicates more inconsistent intensity stability and a greater contribution to coronary artery quality. A smaller difference suggests a greater likelihood that multiple key features are influenced by the same physiological regulatory mechanism, thus requiring a reduction in the influence of strongly correlated features on the contribution to imaging. Therefore, the difference in intensity stability coefficients is positively correlated with the contribution to coronary artery quality, and the intensity stability coefficient itself is positively correlated with the contribution to coronary artery quality.

[0048] In one embodiment of the present invention, the mean difference in intensity stability coefficients between each key imaging feature and different key similar features in the current stenosis level of the patient is obtained. The product of the mean difference and the intensity stability coefficient of each key imaging feature is calculated as the coronary artery quality contribution of each key imaging feature in the current stenosis level of the patient. For the current stenosis level of the patient... The formula for the contribution of coronary artery quality is expressed as: ;in, Indicates the first The first of the narrow levels The contribution of key features in imaging to coronary artery quality; Indicates the first The first of the narrow levels Intensity stability coefficient of key features in image representation; Indicates the first The first of the narrow levels The first key feature of image representation Intensity stability coefficient of key similar features; Indicates the first The number of key similar features that represent key features in an image.

[0049] Step S4: Based on the contribution of key imaging features to coronary artery quality in the current stenosis level of the patient, obtain the coronary artery stenosis evaluation score for the current stenosis level of the patient.

[0050] The contribution of coronary artery quality reflects the independence and stability weight of each key feature in the imaging. The greater the contribution of coronary artery quality, the more stable and independent the key feature in the imaging is, and the more accurately the coronary artery quality can be assessed.

[0051] Preferably, in one embodiment of the present invention, the method for obtaining the coronary artery stenosis evaluation score includes: Based on the contribution of all key imaging features to coronary artery quality in the current stenosis level of the patient, gradient boosting trees are used for pre-training to obtain the basic contribution coefficient of each key imaging feature in the current stenosis level of the patient. Gradient boosting trees can more accurately identify the unique features of the stenosis level, which is beneficial for assessing the quality features of the patient's stenosis level. Gradient boosting trees are a well-known technique in the art and will not be elaborated here.

[0052] The summation of the products between the contribution baseline coefficients of all key imaging features in the current stenosis grade of the patient and the contribution of coronary artery quality is obtained and normalized to serve as the coronary artery stenosis evaluation score for the current stenosis grade of the patient.

[0053] It should be noted that, in the embodiments of the present invention, normalization is performed by linear normalization or a normalization function, for example, by obtaining the maximum and minimum values ​​of the sum of products in different narrowing levels and performing maximum-minimum normalization. The specific means are well known to those skilled in the art and will not be described in detail here.

[0054] Based on this, the higher the coronary artery stenosis evaluation score, the greater the functional assessment of coronary artery stenosis, and the higher the risk of functional ischemia. By analyzing the characteristics of coronary artery stenosis, the quality of the patient's coronary arteries can be more accurately evaluated based on the current stenosis level.

[0055] In summary, this invention obtains the stenosis level for each patient based on the quantified feature values ​​of stenosis degree in multidimensional imaging features; it obtains the key probability of the current patient in each imaging feature based on the distribution of quantified feature values ​​of patients in different imaging features within different stenosis levels, and filters out key imaging features; it obtains the coronary artery quality contribution of each key imaging feature in the current patient's stenosis level based on the distribution of quantified feature values ​​of different patients in each key imaging feature within the current patient's stenosis level; and it obtains the coronary artery stenosis evaluation score for the current patient's stenosis level. This invention improves the accuracy of functional assessment of stenosis by accurately obtaining the coronary artery stenosis evaluation score for the current stenosis level.

[0056] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0057] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for functional assessment of coronary artery stenosis based on angiography radiomics, characterized in that, The method includes: Acquire images of the patient's coronary arteries at different projection angles during end-diastole and obtain the characteristic quantification values ​​of the patient's multidimensional imaging features; Based on the feature quantification value of the stenosis degree of each patient in the multidimensional imaging features, the stenosis level of each patient is obtained; based on the distribution of feature quantification values ​​of patients in different imaging features within different stenosis levels, the key probability of the current patient in each imaging feature is obtained, and key imaging features are screened out. Based on the distribution of feature quantification values ​​of key features in each imaging manifestation of different patients in the current stenosis level, the intensity stability coefficient of each key feature in each imaging manifestation of the current stenosis level is obtained; based on the distribution of intensity stability coefficients of key features in different imaging manifestations in the current stenosis level, the coronary artery quality contribution of each key feature in each imaging manifestation of the current stenosis level is obtained. Based on the contribution of key imaging features to coronary artery quality in the current stenosis level of the patient, a coronary artery stenosis evaluation score is obtained for the current stenosis level of the patient. The methods for obtaining the key possibilities include: The mean value of the feature quantification of each imaging feature for all patients in each stenosis level is obtained as the baseline value of the imaging feature. Based on the difference between the current patient's characteristic quantification value for each imaging feature and the baseline value of the imaging feature corresponding to the previous stenosis level, as well as the degree of fluctuation of the characteristic quantification value, the key possibilities of the current patient for each imaging feature are obtained. The methods for obtaining the key possibilities include: Obtain the degree of fluctuation in the characteristic quantitative values ​​of each imaging feature of historical patients in each stenosis grade; The difference between the current patient’s feature quantification value for each imaging feature and the baseline value of the imaging feature corresponding to the previous stenosis level is calculated and summed up as the result of the fluctuation of feature quantification values ​​for all stenosis levels, which is used as the key probability of the current patient for each imaging feature. The method for obtaining the key features of the image representation includes: If the key probability of an image performance feature is greater than or less than the preset fluctuation threshold, the corresponding image performance feature will be used as the key image performance feature. The method for obtaining the strength stability coefficient includes: Based on the differences between the feature quantification values ​​and corresponding image feature baseline values ​​of key features in each imaging manifestation of different patients in the current stenosis level, as well as the image manifestation feature baseline values, the intensity stability coefficient of each key feature in each imaging manifestation of each stenosis level is obtained. The method for obtaining the strength stability coefficient includes: The mean difference between the feature quantification value and the baseline value of the image manifestation feature of different patients in each image manifestation key feature in the current stenosis level is obtained and negatively correlated to serve as the first stability coefficient; the product of the baseline value of the image manifestation feature and the first stability coefficient is obtained as the intensity stability coefficient of each image manifestation key feature in each stenosis level. The method for obtaining the coronary artery quality contribution includes: Based on the intensity stability coefficient distribution of key imaging features in the current stenosis level of the patient, key similarity features of each key imaging feature are obtained. Based on the differences in intensity stability coefficients between key features of each imaging manifestation and different key similar features in the current stenosis level of the patient, as well as the intensity stability coefficient of each key feature of the imaging manifestation, the coronary artery quality contribution of each key feature of the imaging manifestation in the current stenosis level of the patient is obtained. The difference in intensity stability coefficient is positively correlated with the coronary artery quality contribution, and the intensity stability coefficient is positively correlated with the coronary artery quality contribution. The methods for obtaining the key similarity features include: Obtain the relative distance of the intensity stability coefficients of different key features of image representation, select a preset number of relative distances between all key features of image representation in descending order, and use the corresponding key features of image representation as key similar features.

2. The method for functional assessment of coronary artery stenosis based on angiography radiomics according to claim 1, characterized in that, The method for obtaining the narrowness level includes: If the characteristic quantification value of a patient's stenosis degree falls within the stenosis degree range of a certain preset stenosis level, the corresponding stenosis level of the patient will be set to the preset stenosis level.

3. The method for functional assessment of coronary artery stenosis based on angiography radiomics according to claim 1, characterized in that, The method for obtaining the coronary artery stenosis evaluation score includes: Based on the contribution of all key imaging features to coronary artery quality in the current stenosis level of the patient, a gradient boosting tree is used for pre-training to obtain the basic contribution coefficient of each key imaging feature in the current stenosis level of the patient. The summation of the products between the contribution baseline coefficients of all key imaging features in the current stenosis grade of the patient and the contribution of coronary artery quality is obtained and normalized to serve as the coronary artery stenosis evaluation score for the current stenosis grade of the patient.

Images