CN108451540B - Fractional flow reserve measurement method and device - Google Patents
Fractional flow reserve measurement method and device Download PDFInfo
- Publication number
- CN108451540B CN108451540B CN201710086765.5A CN201710086765A CN108451540B CN 108451540 B CN108451540 B CN 108451540B CN 201710086765 A CN201710086765 A CN 201710086765A CN 108451540 B CN108451540 B CN 108451540B
- Authority
- CN
- China
- Prior art keywords
- blood vessel
- stenosis
- flow reserve
- neural network
- fractional flow
- 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
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/50—Clinical applications
- A61B6/504—Clinical applications involving diagnosis of blood vessels, e.g. by angiography
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/02—Devices for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
- A61B6/03—Computerised tomographs
- A61B6/032—Transmission computed tomography [CT]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/50—Clinical applications
- A61B6/507—Clinical applications involving determination of haemodynamic parameters, e.g. perfusion CT
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
- A61B6/5217—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data extracting a diagnostic or physiological parameter from medical diagnostic data
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
- A61B6/5229—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data combining image data of a patient, e.g. combining a functional image with an anatomical image
Abstract
The invention discloses a fractional flow reserve measuring method and a fractional flow reserve measuring device, wherein the method comprises the following steps: acquiring blood vessel characteristic parameters of coronary arteries of a target patient, wherein the blood vessel characteristic parameters comprise a blood vessel cross section radius at a set position and a stenosis characteristic parameter of a coronary artery blood vessel stenosis region; and inputting the blood vessel characteristic parameters into a pre-established fractional flow reserve measurement model to obtain the fractional flow reserve of the target patient. The invention solves the problems of long time consumption and high calculation requirement on equipment in measuring the fractional flow reserve, and can measure the fractional flow reserve efficiently, accurately and in real time.
Description
Technical Field
The embodiment of the invention relates to the technical field of medical image processing, in particular to a fractional flow reserve measuring method and device.
Background
With the development of social economy and the change of life style of human beings, the morbidity and mortality of coronary artery disease are in a remarkable rising trend, the disease age is in a youthful trend, and the life health of people is seriously threatened, so that the prevention and treatment of coronary artery disease for people are reluctant. Meanwhile, with the rapid development of modern technology and the deep research and exploration of medical science, the prevention, diagnosis, treatment and other aspects of coronary artery disease are all comprehensively developed at present.
The current popular technical means for diagnosing and evaluating coronary artery disease mainly include: coronary angiography, intravascular ultrasound, blood flow reserve fraction and other related technical means and evaluation indexes. Among them, Fractional Flow Reserve (FFR) is a functional index really applied to clinical application, and is an important method for quantitatively and quantitatively evaluating coronary physiological function at a fixed point. Furthermore, FFR has been widely verified to have application value in various lesions, such as single branch coronary lesions, multiple branch lesions, bifurcation lesions, left trunk lesions, diffuse cascade lesions, bifurcation lesions, acute coronary syndrome, critical lesions, evaluation of post-Percutaneous Coronary Intervention (PCI) effects, and guidance of the treatment of Coronary Heart Disease (CHD) in combination with other auxiliary technologies, and especially has wide application in critical lesions, providing scientific basis for CHD optimization treatment, and conforming to the development trend of new-age coronary intervention treatment.
At present, FFR measurement techniques mainly include two types, one is invasive FFR technique, and the other is noninvasive FFR technique based on hemodynamic simulation. However, invasive FFR techniques involve surgical risks, use of vasodilators is toxic, may cause patient allergies, and are expensive; the noninvasive FFR technology based on the hemodynamics simulation has the characteristics of noninvasiveness, economy, strong repeatability, high accuracy, more comprehensive information supply and the like, but has large calculation amount, needs a high-performance computer, consumes long time and cannot meet the requirement of real-time measurement.
Disclosure of Invention
In view of this, the present invention provides a method and an apparatus for measuring fractional flow reserve, so as to solve the problems of long time consumption and high calculation requirement on equipment for measuring fractional flow reserve.
In order to achieve the purpose, the invention adopts the following technical scheme:
in one aspect, an embodiment of the present invention provides a fractional flow reserve measurement method, including:
acquiring blood vessel characteristic parameters of coronary arteries of a target patient, wherein the blood vessel characteristic parameters comprise a blood vessel cross section radius at a set position and a stenosis characteristic parameter of a coronary artery blood vessel stenosis region, and the stenosis characteristic parameter is used for reflecting blood vessel structure information at a blood vessel stenosis position;
and inputting the blood vessel characteristic parameters into a pre-established fractional flow reserve measurement model to obtain the fractional flow reserve of the target patient.
In another aspect, an embodiment of the present invention provides a fractional flow reserve measurement apparatus, including:
the system comprises a blood vessel characteristic acquisition module, a data acquisition module and a data processing module, wherein the blood vessel characteristic acquisition module is used for acquiring blood vessel characteristic parameters of coronary arteries of a target patient, the blood vessel characteristic parameters comprise the radius of a cross section of a blood vessel at a set position and stenosis characteristic parameters of a coronary artery blood vessel stenosis area, and the stenosis characteristic parameters are used for reflecting blood vessel structure information at a blood vessel stenosis position;
the measurement model establishing module is used for establishing a fractional flow reserve measurement model in advance;
and the blood flow reserve fraction measuring module is used for inputting the blood vessel characteristic parameters into a pre-established blood flow reserve fraction measuring model to obtain the blood flow reserve fraction of the target patient.
The invention has the beneficial effects that: according to the fractional flow reserve measurement method and device provided by the embodiment of the invention, the fractional flow reserve measurement model is established in advance, and when a new patient is diagnosed with coronary artery disease, the blood vessel characteristic parameters of the new patient are directly input into the pre-established fractional flow reserve measurement model, so that the fractional flow reserve of coronary artery stenosis of the new patient can be rapidly and accurately output. Compared with the invasive FFR technology, the invention is the noninvasive FFR technology, avoids the operation risk and does not need to use vasodilator; compared with the noninvasive FFR technology based on the hemodynamics simulation, the noninvasive FFR technology based on the hemodynamics simulation has the advantages that the calculation requirement on equipment is low, the calculation can be rapidly carried out only by using a household computer, the consumed time is short, and the requirement on real-time measurement can be met.
Drawings
The above and other features and advantages of the present invention will become more apparent to those of ordinary skill in the art by describing in detail exemplary embodiments thereof with reference to the attached drawings, in which:
fig. 1 is a schematic flow chart of a fractional flow reserve measurement method provided by an embodiment of the present invention;
fig. 2 is a schematic flow chart of another fractional flow reserve measurement method provided by an embodiment of the present invention;
fig. 3 is a schematic flow chart of a method for establishing a fractional flow reserve measurement model according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of a neural network provided by an embodiment of the present invention;
fig. 5 is a block diagram of a fractional flow reserve measurement apparatus according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is further explained by the specific implementation mode in combination with the attached drawings. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Fig. 1 is a schematic flow chart of a fractional flow reserve measurement method according to an embodiment of the present invention. The method is suitable for non-invasively measuring fractional flow reserve, and can be performed by fractional flow reserve measuring means. The apparatus may be implemented by software and/or hardware, and the apparatus may be configured in a computer. As shown in fig. 1, the method includes:
and step 110, acquiring blood vessel characteristic parameters of coronary arteries of the target patient.
The blood vessel characteristic parameters comprise the radius of the cross section of the blood vessel at a set position and the stenosis characteristic parameters of the stenosis region of the coronary artery blood vessel, and the stenosis characteristic parameters are used for reflecting the blood vessel structure information at the stenosis position of the blood vessel.
Optionally, the vessel cross-sectional radius at the set position may be the vessel cross-sectional radius of a plurality of selected slices in the coronary computed tomography image of the target patient; the stenosis characteristic parameters may include a stenosis proximal radius, a stenosis minimum radius, a stenosis distal radius, an entrance length, a minimum radius length, an exit length, and a percent diameter reduction ratio. Wherein the entry length is the length of the vessel centerline from the beginning of the stenosis to the beginning of the minimum radius segment of the stenosis in a direction along the coronary artery away from the heart; the minimum radius length is the length of the vessel centerline from the beginning to the end of the stenotic minimum radius segment; the exit length is the length of the vessel centerline from the end of the segment of minimum radius of the stenosis to the beginning of the stenosis in a direction along the coronary artery away from the heart; the percent diameter reduction ratio can be expressed as:
wherein r issDenotes the minimum radius of stenosis, rpDenotes the normal radius of the vessel, r, closest to the proximal end of the stenosisdRepresents the normal radius of the blood vessel closest to the far end of the stenosis, wherein the area of the stenosis of the blood vessel with the degree of reduction of the inner diameter of more than 10 percent is taken as the inner diameter of the blood vessel wall as a reference, and the reduction range of the inner diameter is taken as the reduction rangeThe degree of the area is not more than 10 percent of the normal area of the blood vessel, and the radius of the blood vessel of the area is the normal radius of the blood vessel.
And 120, inputting the blood vessel characteristic parameters into a pre-established fractional flow reserve measurement model to obtain the fractional flow reserve of the target patient.
Optionally, the present operation may combine all the blood vessel characteristic parameters in step 110 to form a characteristic vector, and input the characteristic vector into a pre-established fractional flow reserve measurement model to directly obtain the fractional flow reserve value of the target patient. The fractional flow reserve measurement model can be constructed based on a neural network. For example, the neural network may be trained using diagnostic data from patients with historical coronary artery disease to obtain a fractional flow reserve measurement model.
Therefore, according to the fractional flow reserve measurement method provided by the embodiment, by establishing the fractional flow reserve measurement model in advance, when a coronary artery disease of a new patient is diagnosed, the blood vessel characteristic parameters of the new patient are directly input into the pre-established fractional flow reserve measurement model, so that the fractional flow reserve of coronary artery stenosis of the new patient can be rapidly and accurately output. Compared with the invasive FFR technology, the invention is the noninvasive FFR technology, avoids the operation risk and does not need to use vasodilator; compared with the noninvasive FFR technology based on the hemodynamics simulation, the noninvasive FFR technology based on the hemodynamics simulation has the advantages that the calculation requirement on equipment is low, the calculation can be rapidly carried out only by using a household computer, the consumed time is short, and the requirement on real-time measurement can be met.
Fig. 2 is a schematic flow chart of another fractional flow reserve measurement method according to an embodiment of the present invention. The embodiment is optimized based on the above embodiment, and the step of obtaining the blood vessel characteristic parameters of the coronary artery of the target patient is optimized as follows: and acquiring blood vessel characteristic parameters according to the coronary artery computed tomography image of the target patient.
As shown in fig. 2, the fractional flow reserve measurement method provided in this embodiment includes:
When the highly collimated X-ray beam is used for cross-section scanning (usually cross section) around a certain part of a human body, part of photons are absorbed, the intensity of the X-ray is attenuated, the unabsorbed photons penetrate through the human body and are absorbed by a detector, then the photons are amplified and converted into electron current, and the electron current is input into an electronic computer as an analog signal to be processed and operated, so that an image is reconstructed. In this embodiment, the computed tomography apparatus scans the coronary artery of the target patient, and the scan data is reconstructed into a computed tomography image of the coronary artery after being processed by the computer. Specifically, the operation may include:
A. from the coronary computed tomography images, vessel cross-sectional radii of a plurality of selected slices are determined.
Alternatively, the vessel cross-sectional radii of each cross-section of the computed tomography image may be determined, and the resulting set of vessel cross-sectional radii is used as part of the vessel characteristic parameters.
B. Stenosis upstream and downstream of a target location point in a coronary vessel is identified from a coronary computed tomography image.
The target position point is each point on the centerline of the coronary artery blood vessel, that is, the operation can sequentially identify the stenosis upstream and downstream of each position point in the coronary artery blood vessel, so that the fractional flow reserve of each position point can be correspondingly obtained, and whether the target patient has coronary lesion or not can be accurately judged. Wherein upstream of the location point represents a segment of the blood vessel that is closer to the heart from the location point and downstream of the location point represents a segment of the blood vessel that is further from the heart from the location point.
As described in the above embodiment, the stenosis region of the blood vessel in which the inner diameter is reduced by more than 10% with reference to the inner diameter of the blood vessel wall is identified, so that the stenosis upstream and downstream of the target site in the coronary blood vessel can be identified.
C. And respectively sequencing the upstream and downstream narrowings of the target position point according to the descending order of the narrowing degree of the narrowings from large to small.
D. At least one of the target stenoses ranked in the top order is selected from stenoses upstream and downstream of the target site, respectively.
Optionally, for one position point, four previously ordered target stenoses can be selected from the upstream stenoses and the downstream stenoses, so as to improve the accuracy of the output structure as much as possible under the condition of ensuring that the calculated amount is small.
E. And (5) extracting stenosis characteristic parameters of the target stenosis.
And step 220, inputting the blood vessel characteristic parameters into a pre-established fractional flow reserve measurement model to obtain the fractional flow reserve of the target patient.
Therefore, the blood vessel characteristic parameters are obtained according to the coronary artery computed tomography image of the target patient, the blood vessel characteristic parameters of the target patient can be conveniently and accurately obtained, and the efficiency of measuring the fractional flow reserve is further improved.
Fig. 3 is a schematic flow chart of a method for establishing a fractional flow reserve measurement model according to an embodiment of the present invention. As shown in fig. 3, the method may include:
The training sample set may include sample vessel characteristic parameters and corresponding sample fractional flow reserve of coronary arteries of a plurality of coronary lesion patients. Optionally, the number of the coronary lesion patients is more than 500, and the sample blood vessel characteristic parameters can be obtained from the coronary artery computed tomography image of the coronary lesion patient, and the structure of the sample blood vessel characteristic parameters is the same as that of the blood vessel characteristic parameters described in the above embodiment; the sample fractional flow reserve may be the invasive fractional flow reserve of each coronary lesion patient.
And step 320, constructing a neural network.
Specifically, an input layer, at least one hidden layer, an activation function, an output layer and initial parameters of each layer of nodes of the neural network may be set, where the input layer includes nodes corresponding to the blood vessel characteristic parameters, the at least one hidden layer includes a plurality of nodes, and the output layer includes one node.
Optionally, the activation function is a Sigmoid function, and the output layer includes a modified linear element, i.e. the activation function is
The output layer includes a linear activation function f (x) max (0, x), i.e., when x < 0, f (x) is 0, and when x ≧ 0, f (x) is x. The initial parameters of each node of each layer can be initialized randomly by a computer. In addition, the hidden layer may be a four-layer hidden layer, as shown in fig. 4, the neural network includes an input layer 10, a four-layer hidden layer 20, and an output layer 30. The number of nodes of each hidden layer in the direction from input to output is 256, 64, 16, and 4, respectively. Optionally, the neural network of this embodiment is a fully-connected neural network, that is, each node of the previous layer is connected to all nodes of the next layer. For the neural network, the relationship between the nth layer and the n +1 th layer may be: x is the number ofn+1＝f(W(n)xn+b(n)) Wherein x isnIs a vector whose ith element represents the value at the ith node of the nth layer; x is the number ofn+1Is a vector whose i-th element represents the value at the i-th node of the (n + 1) -th layer; the function f is the Sigmoid function described above; w(n)And b(n)Is a node parameter, where W(n)Is the weight matrix of the nth layer, b(n)Is the bias vector of the n-th layer, and the neural network is trained to train out W(n)And b(n)The optimum value of (c).
And step 330, training the neural network based on the training sample set.
Illustratively, the sample blood vessel characteristic parameters are used as input of the neural network, the sample blood flow reserve fraction is used as expected output of the neural network, and the neural network is trained repeatedly until the optimal parameters of each node of each layer are obtained.
Specifically, first, the cost function of the whole neural network is defined as:
wherein x is(m)Is the vessel characteristic parameter of the mth patient, and the vessel characteristic parameter obtained in the coronary artery computed tomography image of the ith vessel cross section is x(i)Corresponding to a wound FFR value of y(i)；y(m)Is the invasive FFR value of the mth patient; (x)(m),y(m)) Is a sample of the mth patient; m is the number of samples used for training; m is the mth patient; n is the total number of layers of the neural network; n represents the nth layer of the neural network; snRepresenting the number of nodes of the nth layer in the neural network;
representing the weight value between the ith node of the nth layer of the neural network and the jth node of the (n + 1) th layer; f is a sigmoid function, and f' represents the derivative of f; h (x)(m)) Is x(m)The value obtained after calculation of the whole neural network.
Then, calculating the partial derivative value of the cost function J by using a back propagation algorithm
And
illustratively, given a patient sample (x)(m),y(m)) First, we perform forward propagation operation, calculate activation values of each layer in the neural network, and output value h (x)(m)). Then, for node i of the nth (2 ≦ N < N), we calculate its residual
The residual indicates how much the node has an effect on the residual of the final output value. For the final output node, we can directly calculate the difference between the activation value generated by the neural network and the actual value, and define the difference as
(the Nth layer represents the output layer). For the hidden layer, we willWeighted average calculation using node residuals
Optionally, the activation values of the layers in the neural network may be calculated by using a forward propagation formula, where the forward propagation formula includes:
z(2)＝W(1)x+b(1) (1)
a(n+1)＝f(z(n+1))，n≥1 (2)
z(n+1)＝W(n)a(n)+b(n)，n≥2 (3)
wherein x is an input blood vessel characteristic parameter, a(n)Indicating the activation value of the nth layer. Thus, the activation value of each layer can be obtained from equations (1), (2), and (3).
From this, for the nth layer (output layer) it can be calculated:
δ(N)＝-(y-a(N))*′(z(N)) (4)
wherein y is the output invasive FFR value.
For the nth (2. ltoreq. N < N) layer, it can be calculated:
δ(n)＝((W(n))Tδ(n+1))*f′(z(n)) (5)
with reference to equations (4) and (5), the partial derivative value of the cost function J can be calculated, which is:
and finally, updating each node parameter according to the partial derivative value of the cost function J. Specifically, an all-zero matrix Δ W is set(n)And all-zero vector Δ b(n)Wherein, Δ W(n)And W(n)Of the same dimension,. DELTA.b(n)And b(n)Are the same. After calculating the partial derivative value of the cost function J, the value of Δ W is calculated(n)Is updated to
Will delta b(n)Is updated to
Finally, the updated Δ W is utilized(n)And Δ b(n)Updating node parameters, e.g. W(n)Is updated to
B is to(n)Is updated to
Wherein, the value of alpha can be 0.01, thereby repeatedly training the neural network and finally obtaining the optimal W(n)And b(n)。
And step 340, taking the trained neural network as a fractional flow reserve measurement model.
Through a large number of clinical experiments, the difference value of the FFR value obtained by the method compared with the FFR value output by the existing noninvasive FFR technology (FFR-CT) based on hemodynamics is not more than 0.01, so the consistency of the scheme and the FFR-CT is good.
According to the fractional flow reserve measurement method provided by the embodiment, the fractional flow reserve measurement model is established in advance, and when a new patient is diagnosed with coronary artery disease, the blood vessel characteristic parameters of the new patient are directly input into the pre-established fractional flow reserve measurement model, so that the fractional flow reserve of coronary artery stenosis of the new patient can be rapidly and accurately output. Compared with the invasive FFR technology, the invention is the noninvasive FFR technology, avoids the operation risk and does not need to use vasodilator; compared with the noninvasive FFR technology based on the hemodynamics simulation, the noninvasive FFR technology based on the hemodynamics simulation has the advantages that the calculation requirement on equipment is low, the calculation can be rapidly carried out only by using a household computer, the consumed time is short, and the requirement on real-time measurement can be met.
Fig. 5 is a block diagram of a fractional flow reserve measurement apparatus according to an embodiment of the present invention. As shown in fig. 5, the apparatus includes a blood vessel characteristic acquisition module 1, a measurement model establishment module 2, and a fractional flow reserve measurement module 3.
The blood vessel characteristic acquisition module 1 is used for acquiring blood vessel characteristic parameters of coronary arteries of a target patient, wherein the blood vessel characteristic parameters include a blood vessel cross section radius at a set position and a stenosis characteristic parameter of a coronary artery blood vessel stenosis region, and the stenosis characteristic parameter is used for reflecting blood vessel structure information at a blood vessel stenosis position;
the measurement model establishing module 2 is used for establishing a fractional flow reserve measurement model in advance;
the fractional flow reserve measurement module 3 is used for inputting the blood vessel characteristic parameters into a pre-established fractional flow reserve measurement model to obtain the fractional flow reserve of the target patient.
Optionally, the blood vessel characteristic obtaining module 1 may include:
and the blood vessel characteristic acquisition unit is used for acquiring blood vessel characteristic parameters according to the coronary artery computed tomography image of the target patient.
In this embodiment, the blood vessel characteristic obtaining unit may include:
a vessel cross-section radius measuring subunit for measuring the vessel cross-section radii of the plurality of selected slices from the coronary computed tomography image;
a stenosis identifying subunit, configured to identify a stenosis upstream and downstream of a target location point in a coronary artery blood vessel from the computed tomography image of the coronary artery, wherein the target location point is each point on a centerline of the coronary artery blood vessel;
the narrow sequencing subunit is used for respectively sequencing the upstream and downstream narrow of the target position point according to the sequence of the narrow radius reduction degree from large to small;
a stenosis selecting subunit for selecting at least one target stenosis ranked in the top order from the stenosis upstream and downstream of the target site, respectively;
and the narrow characteristic parameter subunit is used for extracting the narrow characteristic parameters of the target narrow.
Based on the above scheme, the measurement model establishing module 3 of the present embodiment may include:
the training sample set acquisition unit is used for acquiring a training sample set, wherein the training sample set comprises sample blood vessel characteristic parameters of coronary arteries of a plurality of coronary lesion patients and corresponding sample blood flow reserve fractions;
the neural network construction unit is used for constructing a neural network;
the neural network training unit is used for training a neural network based on the training sample set;
and the measurement model determining unit is used for taking the trained neural network as a fractional flow reserve measurement model.
Optionally, the neural network constructing unit may be specifically configured to:
the method comprises the steps of setting an input layer, at least one hidden layer, an activation function, an output layer and initial parameters of nodes of each layer of the neural network, wherein the input layer comprises nodes corresponding to blood vessel characteristic parameters, the at least one hidden layer comprises a plurality of nodes, and the output layer comprises one node.
Wherein, the activation function can be a Sigmoid function, and the output layer can comprise a modified linear unit.
Optionally, the neural network training unit may be specifically configured to:
and (3) taking the characteristic parameters of the sample blood vessels as the input of the neural network, taking the fractional flow reserve of the sample blood vessels as the expected output of the neural network, and repeatedly training the neural network until the optimal parameters of each node of each layer are obtained.
The fractional flow reserve measurement device provided by the embodiment of the invention and the fractional flow reserve measurement method provided by any embodiment of the invention belong to the same inventive concept, can execute the fractional flow reserve measurement method provided by any embodiment of the invention, and have corresponding functions and beneficial effects. For details of the technology not described in detail in this embodiment, reference may be made to the fractional flow reserve measurement method provided in any embodiment of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (5)
1. A fractional flow reserve measurement device, comprising:
the system comprises a blood vessel characteristic acquisition module, a data processing module and a data processing module, wherein the blood vessel characteristic parameters comprise a blood vessel cross section radius at a set position and a stenosis characteristic parameter of a coronary artery blood vessel stenosis region, the stenosis characteristic parameter is used for reflecting blood vessel structure information at a stenosis position, and the stenosis characteristic parameter comprises a stenosis proximal end radius, a stenosis minimum radius, a stenosis distal end radius, an entrance length, a minimum radius length, an exit length and a percentage diameter reduction ratio;
the measurement model establishing module is used for establishing a fractional flow reserve measurement model in advance;
the blood flow reserve fraction measuring module is used for inputting the blood vessel characteristic parameters into a pre-established blood flow reserve fraction measuring model to obtain the blood flow reserve fraction of the target patient;
the blood vessel characteristic acquisition module comprises: a blood vessel characteristic acquisition unit, which is used for acquiring the blood vessel characteristic parameters according to the coronary artery computed tomography image of the target patient;
the blood vessel feature acquisition unit includes:
a vessel cross-section radius determination subunit, configured to determine vessel cross-section radii of a plurality of selected slices from the coronary computed tomography image;
a stenosis identifying subunit for identifying a stenosis upstream and downstream of a target location point in a coronary vessel from the coronary computed tomography image, wherein the target location point is each point on a centerline of the coronary vessel;
the narrow sequencing subunit is used for respectively sequencing the upstream and downstream narrow of the target position point according to the sequence of the narrow radius reduction degree from large to small;
a stenosis selecting subunit for selecting at least one target stenosis ranked in the top order from the stenosis upstream and downstream of the target site, respectively;
and the stenosis characteristic parameter subunit is used for extracting the stenosis characteristic parameters of the target stenosis.
2. The apparatus of claim 1, wherein the measurement model building module comprises:
the training sample set acquisition unit is used for acquiring a training sample set, wherein the training sample set comprises sample blood vessel characteristic parameters of coronary arteries of a plurality of coronary lesion patients and corresponding sample blood flow reserve fractions;
the neural network construction unit is used for constructing a neural network;
a neural network training unit for training the neural network based on the training sample set;
and the measurement model determining unit is used for taking the trained neural network as the fractional flow reserve measurement model.
3. The apparatus according to claim 2, wherein the neural network construction unit is specifically configured to:
setting an input layer, at least one hidden layer, an activation function, an output layer and initial parameters of each layer of nodes of the neural network, wherein the input layer comprises nodes corresponding to the blood vessel characteristic parameters, the at least one hidden layer comprises a plurality of nodes, and the output layer comprises one node.
4. The apparatus of claim 3, wherein the activation function is a Sigmoid function and the output layer comprises a modified linear element.
5. The apparatus of claim 2, wherein the neural network training unit is specifically configured to:
and taking the sample blood vessel characteristic parameters as the input of the neural network, taking the sample blood flow reserve fraction as the expected output of the neural network, and repeatedly training the neural network until the optimal parameters of each node of each layer are obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710086765.5A CN108451540B (en) | 2017-02-17 | 2017-02-17 | Fractional flow reserve measurement method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710086765.5A CN108451540B (en) | 2017-02-17 | 2017-02-17 | Fractional flow reserve measurement method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108451540A CN108451540A (en) | 2018-08-28 |
| CN108451540B true CN108451540B (en) | 2021-08-31 |
Family
ID=63228806
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710086765.5A Active CN108451540B (en) | 2017-02-17 | 2017-02-17 | Fractional flow reserve measurement method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108451540B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109523515A (en) * | 2018-10-18 | 2019-03-26 | 深圳市孙逸仙心血管医院（深圳市心血管病研究所） | The calculation method of blood flow reserve score |
| CN109288537B (en) * | 2018-11-01 | 2022-08-09 | 杭州晟视科技有限公司 | System, method, apparatus and storage medium for assessing fractional flow reserve |
| CN109326354A (en) * | 2018-11-09 | 2019-02-12 | 深圳市孙逸仙心血管医院（深圳市心血管病研究所） | Based on ANN blood flow reserve Score on Prediction method, apparatus, equipment and medium |
| CN109524119B (en) * | 2018-11-09 | 2023-07-04 | 深圳市孙逸仙心血管医院（深圳市心血管病研究所） | GAN-based fractional flow reserve prediction method, device, equipment and medium |
| CN110584696B (en) * | 2019-09-20 | 2023-05-30 | 杭州晟视科技有限公司 | Fractional flow reserve evaluation method and device and storage medium |
| CN111297388B (en) * | 2020-04-03 | 2023-06-20 | 中山大学 | Fractional flow reserve measuring device |
| CN111652881A (en) * | 2020-07-01 | 2020-09-11 | 杭州脉流科技有限公司 | Coronary artery reconstruction and fractional flow reserve calculation method, device and equipment based on deep learning and readable storage medium |
| CN112132203B (en) * | 2020-09-18 | 2023-09-29 | 中山大学 | Fractional flow reserve measurement method and system based on intravascular ultrasound image |
| CN114330161A (en) * | 2021-12-16 | 2022-04-12 | 深圳市阅影科技有限公司 | Generation method and device of Fractional Flow Reserve (FFR) value determination model |
| CN114334160B (en) * | 2021-12-24 | 2023-11-28 | 北京阅影科技有限公司 | Method, device and computer readable storage medium for solving vascular function index |
Family Cites Families (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4421203B2 (en) * | 2003-03-20 | 2010-02-24 | 株式会社東芝 | Luminous structure analysis processing device |
| US7693909B2 (en) * | 2007-04-25 | 2010-04-06 | Siemens Medical Solutions Usa, Inc. | Method for integrating quantitative measurements of imaging systems with reporting applications of medical recording systems |
| CN101991420B (en) * | 2009-08-27 | 2013-01-16 | 上海西门子医疗器械有限公司 | Method and system for positioning suspected blood vessel narrow area |
| CN105190693B (en) * | 2013-03-01 | 2018-04-06 | 波士顿科学国际有限公司 | System and method for the tube chamber border detection in intravascular ultrasound sequence |
| US9974442B2 (en) * | 2013-06-24 | 2018-05-22 | Toshiba Medical Systems Corporation | Method of, and apparatus for, processing volumetric image data |
| CN104424629B (en) * | 2013-08-19 | 2018-01-26 | 深圳先进技术研究院 | A kind of x-ray chest radiograph lung segmentation method and apparatus |
| US9700219B2 (en) * | 2013-10-17 | 2017-07-11 | Siemens Healthcare Gmbh | Method and system for machine learning based assessment of fractional flow reserve |
| KR101578770B1 (en) * | 2013-11-21 | 2015-12-18 | 삼성전자주식회사 | Apparatus por processing a medical image and method for processing a medical image |
| CN103646188A (en) * | 2013-12-27 | 2014-03-19 | 长春工业大学 | Non-invasive diagnostic method of coronary heart disease based on hybrid intelligent algorithm |
| US9595089B2 (en) * | 2014-05-09 | 2017-03-14 | Siemens Healthcare Gmbh | Method and system for non-invasive computation of hemodynamic indices for coronary artery stenosis |
| US9349178B1 (en) * | 2014-11-24 | 2016-05-24 | Siemens Aktiengesellschaft | Synthetic data-driven hemodynamic determination in medical imaging |
| US10349910B2 (en) * | 2015-03-31 | 2019-07-16 | Agency For Science, Technology And Research | Method and apparatus for assessing blood vessel stenosis |
-
2017
- 2017-02-17 CN CN201710086765.5A patent/CN108451540B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108451540A (en) | 2018-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108451540B (en) | Fractional flow reserve measurement method and device | |
| CN108305246B (en) | Device and system for predicting blood flow characteristics based on medical images | |
| CN108038848B (en) | Fast computing method and system based on medical image sequence plaque stability index | |
| CN111134651B (en) | Method, device and system for calculating fractional flow reserve based on intracavity images and computer storage medium | |
| JP6667999B2 (en) | Image processing apparatus, image processing method, and program | |
| CN106659400B (en) | Device for determining a fractional flow reserve value | |
| US8554490B2 (en) | Automatic vascular model generation based on fluid-structure interactions (FSI) | |
| CN108665449B (en) | Deep learning model and device for predicting blood flow characteristics on blood flow vector path | |
| CN109036551B (en) | Coronary artery physiological index relation establishing and applying method and device | |
| JP2017500179A (en) | A method for assessing stenosis severity by stenosis mapping | |
| US9498133B2 (en) | Optoacoustic-ultrasonic system for coregistered functional and morphological imaging of placentas | |
| Zhang et al. | Patient-specific prediction of abdominal aortic aneurysm expansion using Bayesian calibration | |
| CN111956180A (en) | Method for reconstructing photoacoustic endoscopic tomography image | |
| Wilson et al. | Ultrasonic measurement of abdominal aortic aneurysm wall compliance: a reproducibility study | |
| Auricchio et al. | A clinically applicable stochastic approach for noninvasive estimation of aortic stiffness using computed tomography data | |
| Xiao et al. | Estimation of cardiac stroke volume from radial pulse waveform by artificial neural network | |
| Wong et al. | Modelling of blood flow resistance for an atherosclerotic artery with multiple stenoses and poststenotic dilatations | |
| Dai et al. | Three-dimensional modeling and numerical analysis of fractional flow reserve in human coronary arteries | |
| Aghilinejad et al. | Non-invasive pressure-only aortic wave intensity evaluation using hybrid Fourier decomposition-machine learning approach | |
| Potlov et al. | High-precision evaluation of stress-related properties of blood vessel walls using intravascular optical coherence elastography with forward-view probe | |
| US20210290076A1 (en) | System and a method for determining a significance of a stenosis | |
| Wang et al. | Lyapunov exponents for synchronous 12-lead ECG signals | |
| US20160098531A1 (en) | Method for assessing stenosis severity in a lesion tree through stenosis mapping | |
| CN111325827A (en) | Coronary artery in-vitro model experiment method and system | |
| CN115272447B (en) | Multi-modal image-based fractional flow reserve calculation method, device and equipment |
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 |