CN110313990A - A method of prediction bridge vasopermeability model being established based on wall shear stress characteristics of image in bypass surgery - Google Patents
A method of prediction bridge vasopermeability model being established based on wall shear stress characteristics of image in bypass surgery Download PDFInfo
- Publication number
- CN110313990A CN110313990A CN201910563810.0A CN201910563810A CN110313990A CN 110313990 A CN110313990 A CN 110313990A CN 201910563810 A CN201910563810 A CN 201910563810A CN 110313990 A CN110313990 A CN 110313990A
- Authority
- CN
- China
- Prior art keywords
- bridge
- shear stress
- image
- model
- wall shear
- 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.)
- Granted
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
- A61B2034/101—Computer-aided simulation of surgical operations
- A61B2034/105—Modelling of the patient, e.g. for ligaments or bones
Landscapes
- Health & Medical Sciences (AREA)
- Surgery (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Robotics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
A method of prediction bridge vasopermeability model being established based on wall shear stress characteristics of image in bypass surgery, belongs to model foundation field.The instant blood flow waveform of the bridge blood vessel measured in art is subjected to digitized processing, it is enabled to be assigned to bypass surgery threedimensional model as boundary condition, the wall shear stress cloud atlas of stoma site is acquired by finite element method, after the color characteristic and textural characteristics that extract image, dimension-reduction treatment is carried out to feature with principal component analytical method, then the feature after dimensionality reduction is used to construct the prediction model based on support vector machines.This method can be used for that doctor and patient is helped to understand surgical effect, determine the scheme for the tactful or postoperative check further performed the operation.
Description
Technical field:
The invention belongs to the establishing techniques field of model, essence belongs to machine learning field.More particularly to one kind is in heart
The method that prediction bridge vasopermeability model is established based on wall shear stress characteristics of image in bypass surgery.
Background technique:
Coronary bypass surgery (CABG) be treat coronary heart disease a kind of common modus operandi, presently, there are main problem
It is that operation rear axle blood vessel has the risk of failure.According to statistics, vascular grafts after surgery early stage crash rate be 15-30%, 10
Crash rate after year can reach 50%.And 10 years patency rates of arterial bridge blood vessel are 95%, 15 years are 88%.The failure of bridge blood vessel
Afterwards, myocardial ischemia can occur again for patient, can threat to life when serious.Therefore, for each specific patient, how to predict
The permeability of its bridge blood vessel is of great significance.Patient can be helped to avoid the following adverse cardiac events that may occur,
Doctor can be helped to plan the therapeutic scheme of next step.
Currently, common prediction technique is based on the obtained clinical data of Immediacy Index, for example average flow rate, fight
Dynamic index, diastole flow accounting etc., some scholars obtain some new spies based on flow waveform in the way of signal processing
Sign, such as waveform are fourier transformed the ratio of rear its fundamental wave and harmonic wave.But at present not it has been proposed that with blood flow is based on
The feature of mechanics constructs prediction model.
Root is it was found that haemodynamics is to influence the key factor of bridge vasopermeability.The failure of bridge blood vessel is mainly
As caused by atherosis and vascellum endometrial hyperplasia, and undesirable Hemodynamic Factors are considered as their occurrence and development
Most important factor.Particularly with wall shear stress, wall shear stress is too low or excessively high, and wall shear stress gradient is excessively high all to be caused
The permeability of bridge blood vessel reduces.After model is computed, wall shear stress cloud atlas can intuitively reflect the size of wall shear stress with
Gradient.And at bridge vascular anastomosis it is the main generation area of atherosis and endometrial hyperplasia, therefore by the wall at previous anastomotic
Face shearing stress feature accounts for range, is of great significance to the forecasting accuracy of lift scheme.
Summary of the invention:
It is logical that the present invention proposes that one kind establishes prediction bridge blood vessel based on wall shear stress characteristics of image in bypass surgery
The method of permeability model, compared to previous method, the advantage of this method is that introducing more close with penetrating sexual intercourse
Wall shear stress factor, this makes in the case where limited sample size, and the accurate performance of constructed prediction model is mentioned
It rises.
Technical solution is as follows:
It is a kind of that prediction bridge vasopermeability model being established based on wall shear stress characteristics of image in bypass surgery
Method, feature include the following steps:
(1) in art Immediacy Index waveform picture digitized processing；
(2) building of bypass surgery threedimensional model；
(3) in bypass surgery threedimensional model the wall shear stress characteristics of image of stoma site extraction and dimension-reduction treatment；
(4) building of bridge vasopermeability prediction model.
The step of wherein described (1), specifically comprises the following steps:
1.1 intercept out the waveform picture of each bridge blood vessel from the report of postoperative Immediacy Index, pay attention to avoiding as far as possible
The word segment being truncated in report；
Colored waveform picture handle as gray scale picture by 1.2, chooses certain intensity value ranges, will waveform in picture
Background colour and reference axis distinguish；
Abscissa setting start time and end of time in 1.3 pairs of pictures, acquire each pixel on abscissa
The time step of representative；
1.4 pairs of each time steps calculate the pixel number being spaced between waveform and abscissa, while waveform are arranged most
Big value, acquires flow value representated by each longitudinal pixel, can obtain the bridge vascular flow based on time step in this way
Sequence；
1.5 after waveform stabilization, chooses the flow waveform of a cycle, extracts its flow sequence to carry out next step
Work.Such as see Fig. 1
The step of wherein described (2) bypass surgery threedimensional model, design parameter include the following:
The ideal model of 2.1 building bypass surgeries, simulates coronary artery part and bridge vasculature part, bridge blood vessel with straight round tube respectively
The angle intersected with coronary artery is 45 degree；
2.2 bridge blood vessel pipe ranges are set as 120mm, and the pipe range of coronary artery is set as 100mm, and previous anastomotic front end is in coronary artery
10mm, previous anastomotic rear end are 90mm；Front and back is consistent with blood flow direction；
2.3 coronary artery calibers are set as 3mm, the bore of bridge blood vessels caliber ultrasonic probe according to used in Transit-time flowmeter
To be arranged；
2.4 are based on previous anastomotic center, intercept the square that a side length is 15mm, the region in square is will
The region for the previous anastomotic wall shear stress image to be extracted；
The flow waveform of a cycle extracted in step 1.5 is added in bridge blood as inlet flow rate boundary condition by 2.5
Tube inlet, while being outlet pressure boundary condition in the outlet of coronary artery setting 0mmHg.Such as see Fig. 2
The step of wherein described (3), specifically includes as follows:
A cycle is divided equally into 10 moment by 3.1, is cut and is answered in the wall surface that each moment extracts stoma site
The range of power cloud atlas, cloud atlas is each attached to 0-1.2Pa.Such as see Fig. 3
3.2 for each moment wall shear stress cloud atlas, extract its color characteristic and textural characteristics respectively.For face
Color characteristic seeks feature with the method for color grey level histogram.It is R, tri- Color Channels of G, B, to every by picture breakdown
One Color Channel seeks this six kinds of features of its mean value, variance, standard deviation, the degree of bias, kurtosis, energy, entropy.For textural characteristics,
Feature is sought with the method for gray level co-occurrence matrixes.Gray level image is converted by color image, gray level co-occurrence matrixes are set
Four scanning directions are (0,1), and (- 1,1), (- 1,0), (- 1, -1) seeks the gray level co-occurrence matrixes on this four direction.To every
One gray level co-occurrence matrixes seeks this four features of its contrast, correlation, energy, homogeney.To sum up, each when be carved with
34 features at ten moment of each bridge vessel extraction, share 340 features, are arranged successively.
3.3 carry out dimensionality reduction to extracted 340 features with the method for principal component analysis, and last each bridge blood vessel obtains
To 19 features Jing Guo dimension-reduction treatment, this 19 features can indicate 90% or more information content.
The step of wherein described (4) includes:
4.1 construct bridge vasopermeability prediction model using support vector machines, select radial base (RBF) kernel function, seek
Optimal penalty coefficient (C) and kernel function radius (g) are looked for, bridge blood vessel is established with the wall shear stress characteristics of image of stoma site
Permeability prediction model；
When 4.2 penalty coefficients optimal for the searching being previously mentioned in step 2.1 and kernel function radius are to (C, g), then utilize
The method of grid search and cross validation detects the value of each C and g according to certain step-length in a certain range in turn
(above-mentioned to select as needed, not certain limitation), it is right for (C, the g) of the search of each step, it will be rolled over to intersect with k- and tested
The method of card calculates its Average Accuracy, finally takes so that classification accuracy highest (C, g) is right；
4.3 as the bridge blood vessel data of training sample include Immediacy Index data and postoperative actual computer tomography
(CTA) data are scanned, training sample are labeled as by unobstructed or failure according to postoperative CTA data, with the previous anastomotic wall surface after dimensionality reduction
Shearing stress characteristics of image trains the support vector machines, obtains the prediction model of bridge vasopermeability.
The method of the present invention can establish prediction bridge vasopermeability model, the prediction established according to the method for the present invention well
Bridge vasopermeability model has very high accuracy, can be used for that doctor and patient is helped to understand surgical effect, determines further
The scheme of the tactful or postoperative check of operation.
Detailed description of the invention:
Fig. 1: instant blood flow waveform digitizes schematic diagram；
Fig. 2: bypass surgery threedimensional model, Blocked portion is that wall shear stress cloud atlas extracts region in figure；
Fig. 3: the wall shear stress cloud atlas extracted under different moments
Specific embodiment:
Below with reference to embodiment, the present invention will be further described, but the present invention is not limited to following embodiments.
Embodiment 1
To each bridge blood vessel, its waveform is intercepted in instant blood flow waveform picture and digitized processing, root are carried out to it
The bypass surgery threedimensional model for meeting the bridge blood vessel is constructed according to the size of bridge vascular group and ultrasonic probe.Flow will be digitized
Waveform adds to bridge vascular entrance, and the pressure of 0mmHg is added to coronary artery outlet as boundary condition, is counted using finite element software
It calculates.
The characteristics of image of the wall shear stress cloud atlas at bridge vascular anastomosis position is extracted after calculating and carries out dimensionality reduction, is made
It is characterized and constructs prediction model using support vector machines.RBF kernel function is selected, the method pair of grid search and cross validation is utilized
(C, g) coefficient carries out optimizing, chooses the coefficient of optimal (C, g) as prediction model (referring specifically to summary of the invention step).
We are predicted using common Clinical symptoms and wall shear stress characteristics of image respectively as two class of feature construction
Model acquires their Average Accuracies in the case where optimal (C g), average sensitivity and average with the method for cross validation
Specificity, to be used to the performance of comparison prediction model.As a result it see the table below:
As can be known from the results, after using wall shear stress characteristics of image as feature, the accuracy rate of prediction model really can
It is promoted, it is more notable to the promotion of sensitivity.
The method of the present invention has selected 61 bridge blood vessel datas of 37 patients to test, and has collected the art of every bridge blood vessel
In blood flow waveform image data and postoperative CTA review result immediately.According to review result, it is known that there is 21 mistakes in bridge blood vessel
Effect, 40 unobstructed.
Claims (3)
1. a kind of side for establishing prediction bridge vasopermeability model based on wall shear stress characteristics of image in bypass surgery
Method, feature include the following steps:
(1) in bypass surgery threedimensional model the wall shear stress characteristics of image of stoma site extraction and dimension-reduction treatment；
(2) building of bridge vasopermeability prediction model.
2. one kind described in accordance with the claim 1 establishes prediction bridge based on wall shear stress characteristics of image in bypass surgery
The method of vasopermeability model, wherein it is described the step of (1) specifically include it is as follows:
A cycle is divided equally into 10 moment by 1.1, and the wall shear stress cloud of stoma site is extracted at each moment
Figure, the range of cloud atlas are each attached to 0-1.2Pa.
1.2 for each moment wall shear stress cloud atlas, extract its color characteristic and textural characteristics respectively.For color spy
Sign, seeks feature with the method for color grey level histogram.It is R, tri- Color Channels of G, B, to each by picture breakdown
Color Channel seeks this six kinds of features of its mean value, variance, standard deviation, the degree of bias, kurtosis, energy, entropy.For textural characteristics, use
The methods of gray level co-occurrence matrixes seeks feature.Gray level image is converted by color image, four of gray level co-occurrence matrixes are set
Scanning direction is (0,1), and (- 1,1), (- 1,0), (- 1, -1) seeks the gray level co-occurrence matrixes on this four direction.To each
Gray level co-occurrence matrixes seek this four features of its contrast, correlation, energy, homogeney.To sum up, each when be carved with 34
Feature at ten moment of each bridge vessel extraction, shares 340 features, is arranged successively.
1.3 carry out dimensionality reduction to extracted 340 features with the method for principal component analysis, and last each bridge blood vessel obtains
19 features Jing Guo dimension-reduction treatment, this 19 features can indicate 90% or more information content.
3. one kind described in accordance with the claim 1 establishes prediction bridge based on wall shear stress characteristics of image in bypass surgery
The method of vasopermeability model, wherein it is described the step of (2) include:
2.1 construct bridge vasopermeability prediction model using support vector machines, select radial base (RBF) kernel function, find most
Excellent penalty coefficient (C) and kernel function radius (g), establishes bridge vascular permeability with the wall shear stress characteristics of image of stoma site
Property prediction model；
When 2.2 penalty coefficients optimal for the searching being previously mentioned in step 2.1 and kernel function radius are to (C, g), then grid is utilized
The method of search and cross validation, in a certain range, the value for detecting each C and g in turn according to certain step-length is (above-mentioned
Can select as needed, not certain limitation), it is right for (C, the g) of the search of each step, the side of cross validation will be rolled over k-
Method calculates its Average Accuracy, finally takes so that classification accuracy highest (C, g) is right；
2.3 as the bridge blood vessel data of training sample include Immediacy Index data and postoperative actual computer tomoscan
(CTA) training sample is labeled as unobstructed or failure according to postoperative CTA data, is cut and answered with the previous anastomotic wall surface after dimensionality reduction by data
Power characteristics of image trains the support vector machines, obtains the prediction model of bridge vasopermeability.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910563810.0A CN110313990B (en) | 2019-06-26 | 2019-06-26 | Method for establishing prediction bridge vascular permeability model based on wall shear stress image characteristics in heart bypass surgery |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910563810.0A CN110313990B (en) | 2019-06-26 | 2019-06-26 | Method for establishing prediction bridge vascular permeability model based on wall shear stress image characteristics in heart bypass surgery |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110313990A true CN110313990A (en) | 2019-10-11 |
| CN110313990B CN110313990B (en) | 2021-05-14 |
Family
ID=68120500
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910563810.0A Active CN110313990B (en) | 2019-06-26 | 2019-06-26 | Method for establishing prediction bridge vascular permeability model based on wall shear stress image characteristics in heart bypass surgery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110313990B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113609787A (en) * | 2021-08-27 | 2021-11-05 | 中国核动力研究设计院 | Electric valve fault diagnosis method, terminal equipment and readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104116563A (en) * | 2014-07-02 | 2014-10-29 | 北京工业大学 | Hemodynamic fast planning method for personalized coronary artery bypass graft surgery |
| WO2016065161A1 (en) * | 2014-10-22 | 2016-04-28 | Indiana University Research And Technology Corporation | Unified computational method and system for patient-specific hemodynamics |
| DE102014222804A1 (en) * | 2014-11-07 | 2016-05-12 | Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. | Apparatus and method for determining wall shear stress and system for detecting atherosclerosis |
| CN106384114A (en) * | 2016-09-22 | 2017-02-08 | 哈尔滨理工大学 | Support vector machine-based adaptive image binarization method |
| CN109637657A (en) * | 2018-11-27 | 2019-04-16 | 北京工业大学 | The method that bypass surgery based on haemodynamics model establishes prediction bridge vasopermeability model |
| CN109740692A (en) * | 2019-01-13 | 2019-05-10 | 胡燕祝 | A kind of target classifying method of the logistic regression based on principal component analysis |
-
2019
- 2019-06-26 CN CN201910563810.0A patent/CN110313990B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104116563A (en) * | 2014-07-02 | 2014-10-29 | 北京工业大学 | Hemodynamic fast planning method for personalized coronary artery bypass graft surgery |
| WO2016065161A1 (en) * | 2014-10-22 | 2016-04-28 | Indiana University Research And Technology Corporation | Unified computational method and system for patient-specific hemodynamics |
| DE102014222804A1 (en) * | 2014-11-07 | 2016-05-12 | Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. | Apparatus and method for determining wall shear stress and system for detecting atherosclerosis |
| CN106384114A (en) * | 2016-09-22 | 2017-02-08 | 哈尔滨理工大学 | Support vector machine-based adaptive image binarization method |
| CN109637657A (en) * | 2018-11-27 | 2019-04-16 | 北京工业大学 | The method that bypass surgery based on haemodynamics model establishes prediction bridge vasopermeability model |
| CN109740692A (en) * | 2019-01-13 | 2019-05-10 | 胡燕祝 | A kind of target classifying method of the logistic regression based on principal component analysis |
Non-Patent Citations (5)
| Title |
|---|
| WENXIN WANG;BOYAN MAO;HAORAN WANG;XUEYING GENG;XI ZHAO;HUIXIA Z: "Hemodynamic analysis of sequential graft from right coronary system to left coronary system", 《BIOMEDICAL ENGINEERING ONLINE》 * |
| 丁金立，刘有军，王枫: "一种可以避免血管再狭窄的双移植管搭桥方式的数值模拟", 《医学生物力学》 * |
| 李兰兰，刘有军，毛伯，张慧霞: "不同冠脉后负荷下竞争流对左乳内动脉桥血流动力学影响", 《北京生物医学工程》 * |
| 毛伯䶮，刘有军，王文馨，李鲍: "《冠脉搭桥手术移植管即时通透性的血流动力学研究》", 《医用生物力学》 * |
| 赵夕: "冠状动脉搭桥手术规划的血流动力学研究", 《中国博士学位论文全文数据库（电子期刊）》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113609787A (en) * | 2021-08-27 | 2021-11-05 | 中国核动力研究设计院 | Electric valve fault diagnosis method, terminal equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110313990B (en) | 2021-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109637657A (en) | The method that bypass surgery based on haemodynamics model establishes prediction bridge vasopermeability model | |
| US20110229002A1 (en) | Automatic Detection and Quantification of Plaque in the Coronary Arteries of Subjects from CT Scans | |
| Rizzoni et al. | How to assess microvascular structure in humans | |
| CN108765422A (en) | A kind of retinal images blood vessel automatic division method | |
| Friedman | Evaluation of fibrosis and hepatitis C | |
| Atkinson | Shear rate normalization is not essential for removing the dependency of flow-mediated dilation on baseline artery diameter: past research revisited | |
| Kumar et al. | Feature optimized machine learning framework for unbalanced bioassays | |
| CN110313990A (en) | A method of prediction bridge vasopermeability model being established based on wall shear stress characteristics of image in bypass surgery | |
| CN109410267A (en) | A kind of coronary artery segmentation appraisal procedure and system | |
| Varadarajan et al. | Left ventricular diastolic abnormalities in vegetarians compared with non-vegetarians | |
| Cattran | Historical aspects of proteinuria | |
| Yudha et al. | Benefits of Steeping Black Tea As a Negative Contrast Medium on Ct Urography Examination | |
| RU2723626C1 (en) | Method for prediction of clinical course of type 2 diabetes mellitus with application of various types of glucose-lowering therapy on the basis of estimation of glycemia variability | |
| Cirovic et al. | Vortex formation time in female athletes | |
| CN109461139B (en) | Liver cancer quantitative analysis method based on dynamic MRI information fusion | |
| Zytoon et al. | Real-time shear wave elastography for assessing liver fibrosis in patients with chronic hepatitis C | |
| CN117954083B (en) | Preeclampsia prediction method based on machine learning algorithm | |
| CN113420480B (en) | Arterial plaque rupture assessment method and device and storage medium | |
| Drewniak et al. | Methodology and evaluation of the renal arterial system | |
| Baker | Controlled perfusion pressure and vascular volume changes in dog forelimbs | |
| CN107561175A (en) | A kind of evaluation method of glomerular sclerosis rat model | |
| Hand et al. | Prediction of outcome after stroke | |
| Stephens | Machine Learning Predicts Risk of Death in Patients with Suspected or Known Heart Disease | |
| Till et al. | Differential diagnostic results of clinical echography in orbital tumors | |
| CN118213073A (en) | Prognosis prediction system for severe limb ischemia patients treated by cell transplantation |
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 |