CN111657883B - Coronary artery SYNTAX score automatic calculation method and system based on sequence radiography - Google Patents
Coronary artery SYNTAX score automatic calculation method and system based on sequence radiography Download PDFInfo
- Publication number
- CN111657883B CN111657883B CN202010495151.4A CN202010495151A CN111657883B CN 111657883 B CN111657883 B CN 111657883B CN 202010495151 A CN202010495151 A CN 202010495151A CN 111657883 B CN111657883 B CN 111657883B
- Authority
- CN
- China
- Prior art keywords
- coronary
- blood vessel
- coronary artery
- image
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/02007—Evaluating blood vessel condition, e.g. elasticity, compliance
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Physiology (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Heart & Thoracic Surgery (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Biophysics (AREA)
- Geometry (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Mathematical Physics (AREA)
- Fuzzy Systems (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Vascular Medicine (AREA)
- Cardiology (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The coronary artery SYNTAX score automatic calculation method and system based on the sequence contrast can automatically complete the identification of coronary artery anatomical structures, the detection of coronary artery stenosis and the judgment of adverse signs of coronary arteries, finally automatically complete the grading of coronary artery SYNTAX, assist in accurately mastering the pathological changes of the coronary arteries of patients and efficiently complete the diagnosis process of the illness states of the patients. The method comprises the following steps: (1) extracting a relatively complete coronary vessel structure from a key frame of a sequence contrast image; (2) identifying the anatomical structure of the coronary vessel structure, and reserving a coronary section corresponding to a coronary anatomical mode map; (3) guiding by the known coronary segment on the key frame to realize the tracking of the corresponding coronary segment on the sequence contrast image; (4) detecting stenotic lesions on the coronary structures of the sequence; (5) judging whether the stenosis is ill or not on the basis of the detected stenosis; (6) automatic calculation of the SYNTAX score is performed according to the SYNTAX scoring criteria.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a coronary artery SYNTAX score automatic calculation method based on sequence radiography and a coronary artery SYNTAX score automatic calculation system based on sequence radiography.
Background
Coronary artery attached to the surface of heart can cause myocardial ischemia and anoxia, and clinically causes symptoms such as chest pain and chest distress, and is also called coronary heart disease. Coronary heart disease is a well known important death factor worldwide. Clinically, the most coronary insufficiency is caused by the stenosis of the lumen of the blood vessel, so the most conventional diagnosis idea of coronary heart disease is to judge whether the blood vessel of the coronary artery is deficient or not. Clinically, serial coronary angiography is the most direct and objective method for detecting blood vessels, and is also the gold standard used by doctors to judge coronary stenosis. To be able to quantify the condition of patients with coronary heart disease and to provide objective support for the selection of subsequent surgical regimens, the international authoritative cardiology institute issued guidelines to quantify coronary lesions using the SYNTAX scoring criteria. The grading standard is based on the coronary anatomy and the coronary lesion position, the lesion degree and the lesion characteristics of the patient, and realizes the quantitative evaluation of the coronary lesion condition of the patient.
The calculation of the SYNTAX score requires a physician to correspond each coronary vessel in the two-dimensional coronary angiography image with a coronary segment on the standard coronary anatomical pattern map, and needs to make an accurate judgment on the position, structure and degree of each coronary lesion and whether there is an adverse sign. In general, the score calculation process requires a physician to perform three steps of coronary stenosis determination, coronary anatomy identification, and stenosis artifact analysis based on two-dimensional sequence contrast images. However, due to the injection of the contrast agent, the artificial judgment of coronary stenosis is often affected by the transient blockage of the contrast agent in the image, and the judgment of stenosis is erroneous. Meanwhile, due to the perspective imaging principle, more blood vessel overlapping conditions can occur on a two-dimensional angiography image, and structures such as a ring and a pseudo bifurcation point are easily formed, so that the coronary segment is difficult to artificially correspond. In addition, analysis of whether coronary stenosis is a prescribed adverse condition is complicated. For the above reasons, the accuracy of SYNTAX scoring is difficult to guarantee for physicians with inexperienced coronary readings, and even for physicians with abundant experience, the scoring process is also time-consuming and difficult to be widely applied.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides an automatic SYNTAX score calculation method, which automatically completes coronary artery anatomical structure recognition, coronary artery stenosis detection and coronary artery adverse symptom judgment, and finally automatically completes coronary artery SYNTAX scoring, assists doctors to accurately master the coronary artery pathological changes, and efficiently completes disease diagnosis.
The technical scheme of the invention is as follows: the coronary artery SYNTAX score automatic calculation method based on sequence contrast comprises the following steps:
(1) extracting a relatively complete coronary vessel structure from a key frame of a sequence contrast image;
(2) identifying the anatomical structure of the coronary vessel structure, and reserving a coronary section corresponding to a coronary anatomical mode map;
(3) guiding by the known coronary segment on the key frame to realize the tracking of the corresponding coronary segment on the sequence contrast image;
(4) detecting stenotic lesions on the coronary structures of the sequence;
(5) on the basis of the detected stenosis, further judging whether the stenosis has adverse signs;
(6) automatic calculation of the SYNTAX score is performed according to the SYNTAX scoring criteria.
The invention determines a key frame image with complete development of coronary artery structure and less overlapping of blood vessel branches from a sequence contrast image after injection of a contrast agent, extracts the central line of the coronary artery blood vessel on the key frame image to obtain the central line result of all blood vessel structures on the image, then comprehensively utilizes the texture characteristic of the blood vessel structure on the key frame image and the known central line of the blood vessel structure to identify the extracted blood vessel structure in an anatomical segmentation manner, retains the coronary artery segment corresponding to a standard coronary artery anatomical pattern image when the SYNTAX score is clinically calculated, utilizes the coronary artery segment on the key frame image as guidance, tracks and appoints the coronary artery segment on the complete sequence contrast image, then comprehensively analyzes the coronary artery segment structure on the sequence image to complete the coronary artery stenosis detection to obtain a stenosis part, further analyzes the lesion near the stenosis according to 9 adverse sign definitions, and complete SYNTAX score calculation is completed. The invention can automatically complete coronary artery anatomical structure recognition, coronary artery stenosis detection and coronary artery adverse disease sign judgment, and finally automatically complete coronary artery SYNTAX scoring, can assist doctors to master the coronary artery pathological changes of patients, and can efficiently complete the diagnosis process of the disease conditions.
There is also provided a sequence contrast-based coronary SYNTAX score automatic calculation system, comprising:
an extraction module configured to extract a more complete coronary vessel structure from a key frame of a sequence contrast image;
an identification module configured to identify an anatomical structure of a coronary vessel structure and retain a coronary segment corresponding to a coronary anatomical pattern map;
the tracking module is configured to be guided by the known coronary artery segment on the key frame to realize the tracking of the corresponding coronary artery segment on the sequence contrast image;
a detection module configured to perform a detection of a stenosis on a coronary structure of the sequence;
a judging module configured to further judge whether there is a bad symptom on the basis of the detected stenosis;
a calculation module configured to perform an automatic calculation of the SYNTAX score according to a SYNTAX scoring criterion.
Drawings
Fig. 1 is a flowchart of a coronary SYNTAX score automatic calculation method based on sequence contrast according to the present invention.
FIG. 2 is a flow chart of step (1) according to the present invention.
Fig. 3 is a flow chart of step (2) according to the present invention.
Fig. 4 is a flowchart of step (3) according to the present invention.
Fig. 5 is a flowchart of step (4) according to the present invention.
Fig. 6 is a flowchart of a coronary artery adverse condition judgment according to step (5) of the present invention.
Fig. 7 is a flowchart of another coronary artery adverse condition judgment according to step (5) of the present invention.
Fig. 8 is a flowchart of another coronary artery adverse condition judgment according to step (5) of the present invention.
Fig. 9 is a flowchart of another coronary artery adverse condition judgment according to step (5) of the present invention.
Fig. 10 is a flowchart of another coronary artery adverse condition judgment according to step (5) of the present invention.
Detailed Description
As shown in fig. 1, the method for automatically calculating coronary artery SYNTAX score based on sequence contrast comprises the following steps:
(1) extracting a relatively complete coronary vessel structure from a key frame of a sequence contrast image;
(2) identifying the anatomical structure of the coronary vessel structure, and reserving a coronary section corresponding to a coronary anatomical mode map;
(3) guiding by the known coronary segment on the key frame to realize the tracking of the corresponding coronary segment on the sequence contrast image;
(4) detecting stenotic lesions on the coronary structures of the sequence;
(5) on the basis of the detected stenosis, further judging whether the stenosis has adverse signs;
(6) automatic calculation of the SYNTAX score is performed according to the SYNTAX scoring criteria.
The invention determines a key frame image with complete development of coronary artery structure and less overlapping of blood vessel branches from a sequence contrast image after injection of a contrast agent, extracts the central line of the coronary artery blood vessel on the key frame image to obtain the central line result of all blood vessel structures on the image, then comprehensively utilizes the texture characteristic of the blood vessel structure on the key frame image and the known central line of the blood vessel structure to identify the extracted blood vessel structure in an anatomical segmentation manner, retains the coronary artery segment corresponding to a standard coronary artery anatomical pattern image when the SYNTAX score is clinically calculated, utilizes the coronary artery segment on the key frame image as guidance, tracks and appoints the coronary artery segment on the complete sequence contrast image, then comprehensively analyzes the coronary artery segment structure on the sequence image to complete the coronary artery stenosis detection to obtain a stenosis part, further analyzes the stenosis nearby according to 9 adverse sign definitions, the invention can automatically complete coronary artery anatomical structure recognition, coronary artery stenosis detection and coronary artery adverse disease sign judgment, and finally automatically complete coronary artery SYNTAX scoring, thereby assisting doctors to master the coronary artery pathological changes of patients and efficiently completing the diagnosis process of the disease conditions.
Preferably, in the step (1), a coronary vessel region in the key frame contrast image is obtained by a vessel segmentation method, and then the region is subjected to central position enhancement, which is skeleton enhancement; obtaining a blood vessel central line structure with maximum response after the framework in the coronary region is enhanced by using non-maximum inhibition operation; utilizing the initial blood vessel centerline skeleton information, and decomposing the feature value based on tensor to obtain the connection probability of the fracture part in the skeleton; meanwhile, texture information of an original image of the key frame contrast image and coronary artery vessel trend direction information analyzed from the contrast image are utilized to construct connection probability based on texture and direction; the three connection probability maps form an initial blood vessel skeleton bridge-cut connection probability map through weighting fusion; and then, connecting the broken bridges by utilizing path search at the initial vascular skeleton fracture part based on the connection probability map to obtain a complete blood vessel center line result.
Preferably, the step (2) comprises the following substeps:
(2.1) based on the key frame contrast image and the corresponding blood vessel skeleton, encoding the texture features of the blood vessel points in the key frame contrast image by using a convolutional neural network;
(2.2) encoding topological features of the vessel points using a graph embedding operation based on the vessel skeleton;
(2.3) fusing the texture features and the topological features of the blood vessel points to obtain the spatial features of the blood vessel points;
(2.4) decoding the spatial features of the blood vessel points by using a neural network deconvolution module, and recovering the positions of the blood vessel points to the initial image;
(2.5) classifying the anatomical segment for each vessel point in the image using a classification network.
Preferably, in the step (3), using the specified coronary anatomy segment structure obtained from the key frame as a reference, the coronary anatomy segment corresponding to the specified coronary anatomy segment is tracked on the full slice image of the contrast agent overflow in the whole sequence of contrast images.
Preferably, the step (4) comprises the following substeps:
(4.1) selecting a frame image to be detected and 2 continuous frames of images before and after the frame image from the sequence contrast images, wherein the five frames of images form the input of the module, and the characteristic images of the five frames of images are obtained by utilizing an image characteristic extraction network with an attention mechanism;
(4.2) fusing the feature maps obtained at the five different moments in each scale by using a feature fusion network with an attention mechanism;
(4.3) performing re-fusion on the features under different scales by using a common feature pyramid operation in target detection to obtain a final feature map to be processed;
(4.4) consistent with the target detection task, obtaining a plurality of target positioning frames on the basis of the feature map by using the region generation network, then obtaining target frame internal features from the feature map by using the region-of-interest feature extraction network, and classifying the target internal features by using the classification network to obtain the probability value of whether the target internal features are in a narrow category or not; the target frame position coordinates and the width and height are regressed using a regression network.
Preferably, the determination of adverse disease symptoms in step (5) further comprises: open lesions, trifurcate/bifurcation lesions, long lesions, severe flexion lesions;
three-dimensional reconstruction is carried out on the blood vessel in the neighborhood of the narrow part, the specified quantitative index is strictly calculated, and whether the adverse symptoms appear or not is judged according to the definition; judging the opening lesion, and calculating the distance between the stenosis position and the coronary artery opening position; judging trifurcate/bifurcation lesion, calculating the distance between a stenosis position and an opening position of a bifurcation area, and judging the state of the stenosis on each section of the bifurcation; and (3) judging long lesions, calculating the length of the stenosis, and judging severe bending, wherein the curvature of each point of the proximal vessel section of the stenosis part needs to be calculated.
Preferably, the determination of adverse disease symptoms in step (5) further comprises: calcified lesions, thrombotic lesions, total occlusive lesions, diffuse lesions;
calcified lesion judgment, namely, on an angiogram cover sheet, carrying out image enhancement on a coronary artery section region where the stenosis is located, and setting a threshold value to judge whether calcification occurs or not, wherein the coronary artery section region where the stenosis is located is an expansion region corresponding to a coronary artery section on the filling sheet;
judging thrombus lesion, namely enhancing the coronary artery section on the filling sheet around the narrow coronary artery section, and then judging whether thrombus appears on the blood vessel section by using circular detection;
judging total occlusion lesion, analyzing the dissected blood vessel section of the full filling piece of the contrast agent in the sequence contrast image, and if the branch of the main blood vessel section is obviously less than the normal coronary artery, indicating that the patient has total occlusion;
and judging diffuse lesion, namely calculating the radius of a blood vessel section where the stenotic lesion is and the radius of the blood vessel section in the neighborhood of the stenotic lesion, and judging that the patient has the diffuse lesion if a large-area small radius appears.
It will be understood by those skilled in the art that all or part of the steps in the method of the above embodiments may be implemented by hardware instructions related to a program, the program may be stored in a computer-readable storage medium, and when executed, the program includes the steps of the method of the above embodiments, and the storage medium may be: ROM/RAM, magnetic disks, optical disks, memory cards, and the like. Therefore, in accordance with the method of the present invention, the present invention also includes a coronary SYNTAX score automatic calculation system based on sequence contrast, which is generally represented in the form of functional blocks corresponding to the steps of the method. The system comprises:
an extraction module configured to extract a more complete coronary vessel structure from a key frame of a sequence contrast image;
an identification module configured to identify an anatomical structure of a coronary vessel structure and retain a coronary segment corresponding to a coronary anatomical pattern map;
the tracking module is configured to be guided by the known coronary artery segment on the key frame to realize the tracking of the corresponding coronary artery segment on the sequence contrast image;
a detection module configured to perform a detection of a stenosis on a coronary structure of the sequence;
a judging module configured to further judge whether there is a bad symptom on the basis of the detected stenosis;
a calculation module configured to perform an automatic calculation of the SYNTAX score according to a SYNTAX scoring criterion.
Preferably, the identification module performs:
based on the key frame contrast image and the corresponding blood vessel skeleton, encoding the texture characteristics of the blood vessel points by using a convolution neural network;
based on the blood vessel skeleton, encoding the topological characteristics of the blood vessel points by using a graph embedding operation;
fusing the texture features and the topological features of the blood vessel points to obtain the spatial features of the blood vessel points;
decoding the spatial features of the blood vessel points by using a neural network deconvolution module, and recovering the positions of the blood vessel points to the initial image;
a classification network is used to classify the anatomical segment for each vessel point in the image.
Preferably, the detection module performs:
selecting a frame image to be detected and 2 continuous frames of images before and after the frame image from the sequence contrast images, wherein the five frames of images form the input of the module, and obtaining respective characteristic images of the five frames of images by utilizing an image characteristic extraction network with an attention mechanism;
fusing all scales of the feature maps obtained at the five different moments by using a feature fusion network with an attention mechanism;
performing re-fusion on the features under different scales by using a common feature pyramid operation in target detection to obtain a final feature map to be processed;
consistent with the target detection task, obtaining a plurality of target positioning frames on the basis of the feature map by using the region generation network, then obtaining the internal features of the target frames from the feature map by using the region-of-interest feature extraction network, and classifying the internal features of the targets by using the classification network to obtain the probability value of whether the internal features are in a narrow category or not; the target frame position coordinates and the width and height are regressed using a regression network.
One embodiment of the present invention is explained below. The input of the method and system of the invention is a sequence coronary angiography image of a coronary heart disease patient.
Firstly, determining a key frame image in which the heart is in a diastolic state after the injection of the contrast agent and the coronary structure is completely developed and the blood vessel branches are less overlapped from the sequence contrast images. And then extracting the center line of the coronary vessel on the key frame image to obtain the center line result of all the vessel structures on the image. And then, comprehensively utilizing texture features of the blood vessel structure on the key frame image and the known center line of the blood vessel structure, carrying out anatomical segmentation identification on the extracted blood vessel structure, and reserving the coronary segment corresponding to the standard coronary anatomical pattern image when the SYNTAX score is clinically calculated. Then, the coronary segment on the key frame image is used as a guide, and the designated coronary segment is tracked on the complete sequence contrast image. And then comprehensively analyzing the coronary segment structure on the sequence image to finish coronary stenosis detection and obtain a stenosis part. Further analysis of the 9 adverse signs was performed near the stenosis, according to their definition, to complete the complete SYNTAX score calculation. The output of the invention is the coronary SYNTAX score of the patient.
Fig. 2 is an illustration of a prior patent of the inventor (patent application No. cn201711339539.x, title of the invention: a method and system for extracting a centerline of a tubular structure). Firstly, a coronary vessel region in a key frame contrast image is obtained by using a vessel segmentation method, and then the region is subjected to central position enhancement, namely skeleton enhancement. And obtaining the blood vessel central line structure with the maximum response after the framework in the coronary artery region is enhanced by using non-maximum inhibition operation. And (4) utilizing the initial blood vessel centerline skeleton information, and decomposing the feature values based on the tensor to obtain the connection probability of the fracture part in the skeleton. Meanwhile, the method utilizes texture information of an original image of the key frame contrast image and coronary artery vessel trend direction information analyzed from the contrast image to construct connection probability based on texture and direction. And the three connection probability maps form an initial blood vessel skeleton bridge-cut connection probability map through weighting fusion. And then, connecting the broken bridges by utilizing path search at the initial vascular skeleton fracture part based on the connection probability map to obtain a complete blood vessel center line result.
The vessel anatomy identification on the key frame contrast image of the present invention is described in detail as shown in fig. 3.
The method comprises the following steps: based on the key frame contrast image and the corresponding blood vessel skeleton, encoding the texture characteristics of the blood vessel points by using a convolution neural network;
step two: based on the blood vessel skeleton, encoding the topological characteristics of the blood vessel points by using a graph embedding operation;
step three: fusing the texture features and the topological features of the blood vessel points to obtain the spatial features of the blood vessel points;
step four: decoding the spatial features of the blood vessel points by using a neural network deconvolution module, and recovering the positions of the blood vessel points to the initial image;
step five: a classification network is used to classify the anatomical segment for each vessel point in the image.
Fig. 4 is an illustration of a prior patent of the inventor (patent application No. CN201910182030.1, title of the invention: method and apparatus for tracking tubular structures in X-ray contrast image sequences). Using the specified coronary anatomy segment structure derived from the keyframe, the coronary anatomy segment corresponding thereto is tracked over the full slice of contrast-flooded images in the entire sequence of contrast images, with reference thereto.
The detection of coronary stenosis in the sequence contrast images of the present invention is described with reference to fig. 5.
The method comprises the following steps: firstly, a frame image to be detected and 2 continuous frames of images before and after the frame image are selected from the sequence contrast images, and the five frames of images form the input of the cost module. And obtaining the characteristic maps of the five frames of images by using an image characteristic extraction network with an attention mechanism.
Step two: and (4) fusing the feature maps obtained at the five different moments by using a feature fusion network with an attention mechanism at each scale.
Step three: and performing secondary fusion on the features under different scales by using the common feature pyramid operation in the target detection to obtain a final feature map to be processed.
Step four: and according with the target detection task, obtaining a plurality of target positioning frames on the basis of the feature map by utilizing a region generation network, then obtaining the features in the target frames from the feature map by utilizing an interested region feature extraction network, classifying the features in the targets by utilizing a classification network to obtain the probability value of whether the features are in a narrow category or not, and regressing the position coordinates, the width and the height of the target frames by utilizing a regression network. .
FIG. 6 is a flow chart of the present invention for determining the symptoms of coronary artery disease (open lesion, trifurcate/bifurcation lesion, long lesion, severe tortuosity lesion).
Because the open lesion, the trifurcate lesion, the bifurcation lesion, the long lesion and the severe bending lesion are defined by quantitative information such as the length, the position, the curvature of the adjacent blood vessel and the like around the stenosis, the invention carries out three-dimensional reconstruction on the blood vessel in the neighborhood of the stenosis part, strictly calculates the specified quantitative index and judges whether the adverse disease symptoms appear or not according to the definition. Wherein, the judgment of the opening lesion needs to calculate the distance between the stenosis position and the coronary artery opening position. Trigeminal/bifurcation lesion determination requires calculation of the distance of the stenosis location from the opening location of the bifurcation area and discussion of the state of the presence of the stenosis on the various segments of the bifurcation. The judgment of long lesions requires the calculation of the stenosis length, and the judgment of severe bending requires the calculation of the curvature of each point of the proximal vessel segment of the stenosis site.
FIG. 7 is a flowchart of the determination of coronary artery disease symptoms (calcified lesions) according to the present invention. Since calcified lesions appear in the vessel segment where the stenosis is located, and by definition, on the mask where the contrast agent is not injected, the dark image of the vessel region is calcified. Therefore, in the invention, on the contrast mask, the image enhancement is carried out on the coronary section area (the expansion area corresponding to the coronary section on the filling sheet) where the stenosis is positioned, and the threshold value is set to judge whether the calcification appears.
FIG. 8 is a flowchart of the determination of coronary artery disease symptoms (thrombotic lesions) according to the present invention. Similar to calcification judgment, thrombus judgment still surrounds a coronary artery section with stenosis, the coronary artery section on the filling sheet is strengthened by the method, and then circular detection is used for judging whether thrombus appears on the blood vessel section.
Fig. 9 is a flowchart of judging the coronary artery adverse disease (total occlusion lesion) according to the present invention. The invention analyzes the dissected blood vessel segment of the full filling piece of the contrast agent in the sequence contrast image, if the branch of the main blood vessel segment is obviously less than the normal coronary artery, the patient has the complete occlusion.
Fig. 10 is a flowchart of the determination of coronary artery disease symptoms (diffuse lesions) according to the present invention. The diffuse lesion is defined as the condition that the length of the stenosis accounts for more than 75% of the length of the blood vessel section in the coronary section where the stenosis is located and the adjacent coronary section. Therefore, the invention calculates the radius of the blood vessel section where the stenotic lesion is located and the blood vessel section in the neighborhood thereof, and if a large area of small radius appears, the patient is judged to have diffuse lesion.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent variations and modifications made to the above embodiment according to the technical spirit of the present invention still belong to the protection scope of the technical solution of the present invention.
Claims (2)
1. A coronary artery SYNTAX score automatic calculation system based on sequence radiography is characterized in that: it includes:
an extraction module configured to extract a more complete coronary vessel structure from a key frame of a sequence contrast image;
an identification module configured to identify an anatomical structure of a coronary vessel structure and retain a coronary segment corresponding to a coronary anatomical pattern map;
the tracking module is configured to be guided by the known coronary artery segment on the key frame to realize the tracking of the corresponding coronary artery segment on the sequence contrast image;
a detection module configured to perform a detection of a stenosis on a coronary structure of the sequence;
a judging module configured to further judge whether there is a bad symptom on the basis of the detected stenosis;
a calculation module configured to perform an automatic calculation of the SYNTAX score according to a SYNTAX scoring criterion;
the identification module performs:
based on the key frame contrast image and the corresponding blood vessel skeleton, encoding the texture characteristics of the blood vessel points by using a convolution neural network;
based on the blood vessel skeleton, encoding the topological characteristics of the blood vessel points by using a graph embedding operation;
fusing the texture features and the topological features of the blood vessel points to obtain the spatial features of the blood vessel points;
decoding the spatial features of the blood vessel points by using a neural network deconvolution module, and recovering the positions of the blood vessel points to the initial image;
a classification network is used to classify the anatomical segment for each vessel point in the image.
2. The system for automatically calculating coronary SYNTAX score based on sequence contrast according to claim 1, wherein: the detection module performs:
selecting a frame image to be detected and 2 continuous frames of images before and after the frame image from the sequence contrast images, wherein the five frames of images form the input of the module, and obtaining respective characteristic images of the five frames of images by utilizing an image characteristic extraction network with an attention mechanism;
fusing the feature maps of the five frames of images in each scale by using a feature fusion network with an attention mechanism;
performing re-fusion on the features under different scales by using a common feature pyramid operation in target detection to obtain a final feature map to be processed;
consistent with the target detection task, obtaining a plurality of target positioning frames on the basis of the feature map by using a region generation network, then obtaining the internal features of the target frames from the feature map by using an interested region feature extraction network, and classifying the internal features of the target by using a classification network to obtain the probability value of the target belonging to a narrow category; the target frame position coordinates and the width and height are regressed using a regression network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010495151.4A CN111657883B (en) | 2020-06-03 | 2020-06-03 | Coronary artery SYNTAX score automatic calculation method and system based on sequence radiography |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010495151.4A CN111657883B (en) | 2020-06-03 | 2020-06-03 | Coronary artery SYNTAX score automatic calculation method and system based on sequence radiography |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111657883A CN111657883A (en) | 2020-09-15 |
| CN111657883B true CN111657883B (en) | 2021-05-04 |
Family
ID=72385611
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010495151.4A Active CN111657883B (en) | 2020-06-03 | 2020-06-03 | Coronary artery SYNTAX score automatic calculation method and system based on sequence radiography |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111657883B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112837306B (en) * | 2021-02-20 | 2022-11-22 | 薛竟宜 | Coronary artery disease lesion functional quantitative method based on deep learning and mesopic theory |
| CN114418977A (en) * | 2022-01-07 | 2022-04-29 | 乐普（北京）医疗器械股份有限公司 | Method and device for coronary angiography quantitative analysis based on angiography video |
| CN116523813A (en) * | 2022-01-20 | 2023-08-01 | 乐普（北京）医疗器械股份有限公司 | Processing method and device for analyzing fractional flow reserve based on angiographic image |
| CN116188457B (en) * | 2023-04-18 | 2023-07-21 | 天津恒宇医疗科技有限公司 | Processing method and processing system of coronary angiography skeleton diagram |
| CN116704149B (en) * | 2023-08-03 | 2023-11-14 | 杭州脉流科技有限公司 | Method and device for obtaining intracranial arterial stenosis parameters |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108830155A (en) * | 2018-05-10 | 2018-11-16 | 北京红云智胜科技有限公司 | A kind of heart coronary artery segmentation and knowledge method for distinguishing based on deep learning |
| CN109978915A (en) * | 2019-03-11 | 2019-07-05 | 北京理工大学 | The tracking and device of tubular structure in X-ray angiographic image series |
| CN110546646A (en) * | 2017-03-24 | 2019-12-06 | 帕伊医疗成像有限公司 | method and system for assessing vascular occlusion based on machine learning |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957066B (en) * | 2016-04-22 | 2019-06-25 | 北京理工大学 | CT image liver segmentation method and system based on automatic context model |
| US10163209B2 (en) * | 2016-11-23 | 2018-12-25 | Toshiba Medical Systems Corporation | Medical image processing apparatus, medical image processing method, and X-ray CT apparatus |
| CN108198184B (en) * | 2018-01-09 | 2020-05-05 | 北京理工大学 | Method and system for vessel segmentation in contrast images |
| CN111091589B (en) * | 2019-11-25 | 2023-11-17 | 北京理工大学 | Ultrasonic and nuclear magnetic image registration method and device based on multi-scale supervised learning |
-
2020
- 2020-06-03 CN CN202010495151.4A patent/CN111657883B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110546646A (en) * | 2017-03-24 | 2019-12-06 | 帕伊医疗成像有限公司 | method and system for assessing vascular occlusion based on machine learning |
| CN108830155A (en) * | 2018-05-10 | 2018-11-16 | 北京红云智胜科技有限公司 | A kind of heart coronary artery segmentation and knowledge method for distinguishing based on deep learning |
| CN109978915A (en) * | 2019-03-11 | 2019-07-05 | 北京理工大学 | The tracking and device of tubular structure in X-ray angiographic image series |
Non-Patent Citations (1)
| Title |
|---|
| 冠状动脉病变评分系统设计与实现;揭海军;《中国优秀硕士学位论文全文数据库 医药卫生科技辑》;20160215;第62-84页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111657883A (en) | 2020-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111657883B (en) | Coronary artery SYNTAX score automatic calculation method and system based on sequence radiography | |
| Nasr-Esfahani et al. | Segmentation of vessels in angiograms using convolutional neural networks | |
| JP2023138502A (en) | X-ray image feature detection and registration systems and methods | |
| Wu et al. | Automatic detection of coronary artery stenosis by convolutional neural network with temporal constraint | |
| CN110222759B (en) | Automatic identification system for vulnerable plaque of coronary artery | |
| Manniesing et al. | Vessel axis tracking using topology constrained surface evolution | |
| CN110490040B (en) | Method for identifying local vascular stenosis degree in DSA coronary artery image | |
| JP2019193808A (en) | Diagnostically useful results in real time | |
| US9084531B2 (en) | Providing real-time marker detection for a stent in medical imaging | |
| Hao et al. | Sequential vessel segmentation via deep channel attention network | |
| US20220335616A1 (en) | Longitudinal Display Of Coronary Artery Calcium Burden | |
| JP2008534109A (en) | Apparatus and method for positioning a device within a tubular organ | |
| JPH07143977A (en) | Digital image processing method for automatic detection of constriction | |
| JP2006508772A (en) | Medical observation system and method for detecting a boundary of interest in a noisy image | |
| US20220284583A1 (en) | Computerised tomography image processing | |
| JP2011517986A (en) | Automatic detection and accurate segmentation of abdominal aortic aneurysms | |
| CN109478327A (en) | Method for detecting systemic arterial automatically in any kampometer calculation tomographic angiography (CTA) | |
| CN112862833A (en) | Blood vessel segmentation method, electronic device and storage medium | |
| Klein et al. | Automatic segmentation of the wire frame of stent grafts from CT data | |
| Zhao et al. | Automatic aortic dissection centerline extraction via morphology-guided CRN tracker | |
| Zhao et al. | A new approach to extracting coronary arteries and detecting stenosis in invasive coronary angiograms | |
| CN115409758A (en) | Method for comprehensively and quantitatively evaluating angiostenosis from contrast images | |
| M’hiri et al. | Automatic evaluation of vessel diameter variation from 2D X-ray angiography | |
| Markiewicz et al. | Computerized system for quantitative assessment of atherosclerotic plaques in the femoral and iliac arteries visualized by multislice computed tomography | |
| Gao | Automated planning approaches for non-invasive cardiac valve replacement |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |