4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images

Abstract

The present invention discloses a 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images. PC-MRI images are input through an image sequence input port, and the images are classified and verified. The image processing module further extracts vascular structure parameters and blood flow velocity characteristic values ​​through the amplitude image sequence and the phase image sequence. The pulmonary artery pressure calculation and display module then introduces a linear regression model to calculate the mean pulmonary artery pressure; the pulmonary hypertension diagnosis and evaluation module introduces a logistic regression model to diagnose pulmonary hypertension. The progress of pulmonary hypertension is dynamically evaluated based on multiple average pulmonary artery measurement results. The present invention can realize non-invasive pulmonary artery pressure measurement, thereby realizing pulmonary hypertension diagnosis, and evaluating the prognosis of pulmonary hypertension.

Description

Technical Field

[0001] The present invention belongs to the technical field of non-invasive detection of pulmonary artery pressure, and relates to a pulmonary artery pressure measurement and evaluation system, specifically, a 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images. Background Art

[0002] Pulmonary hypertension (PH) is a pathological state of elevated blood pressure in the pulmonary artery. The current gold standard for diagnosing and evaluating the course of PH disease is right heart catheter interventional pressure measurement. Since this method is an interventional measurement method, it cannot be used frequently, and cannot achieve dynamic evaluation of PH, and cannot meet the needs of PH treatment prognosis and risk assessment of acute PH attacks. In the article "Magnetic Resonance-Derived 3-Dimensional Blood Flow and aMeasure of Elevated Means Pulmonary Arterial Pressure" (Circ. Cardiovascular Imaging, 2008), a method for detecting and distinguishing PH by non-invasive phase contrast magnetic resonance imaging (PC-MRI) is disclosed. According to the flow pattern in the pulmonary artery, the normal state, potential PH state and explicit PH state are distinguished. The existence of vortices along the mainstream direction in the main pulmonary artery, the diastolic streamline of blood along the anterior wall of the main pulmonary artery, and the blood flow velocity distribution in the anterior-posterior direction above the pulmonary valve were investigated. Figure 3 These characteristics are based on manual measurements on PC-MRI images, and their clinical application relies on a lot of tedious and time-consuming manual measurements, which limits their popularization and application.

[0003] In addition, the above method only considers the blood flow velocity information provided in the PC-MRI image, which makes the accuracy of PH diagnosis low; this method requires the identification of structures such as the anterior wall of the pulmonary artery and the pulmonary valve, which increases the spatial resolution requirements of PC-MRI imaging and the difficulty of measuring local features; PH discrimination mainly relies on the accurate measurement of some phases in the cardiac cycle, which increases the requirements for the temporal resolution of PC-MRI imaging. The high temporal resolution of PC-MRI imaging will directly lead to an increase in imaging time, limiting the possibility of clinical use. Summary of the invention

[0004] In view of the above problems, the present invention proposes a 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images. It is a system for measuring and estimating pulmonary artery pressure from phase contrast magnetic resonance images (PC-MRI), which can be used for non-invasive diagnosis and dynamic evaluation of pulmonary hypertension, for prognosis evaluation of PH treatment, and for risk prediction of acute PH attacks. Based on phase contrast magnetic resonance images, the present invention studies the vascular structure information and blood flow velocity information related to PH contained in the images, and proposes a method for automatically measuring this information. The linear regression model is used to comprehensively utilize this information to achieve accurate measurement of pulmonary artery pressure. The present invention also uses a logistic regression model to achieve non-invasive diagnosis of PH. Since the method can measure the pulmonary artery non-invasively, simply and automatically, it can be repeated multiple times, which is convenient for long-term monitoring, and can provide a technical basis for the prognosis and dynamic evaluation of PH.

[0005] The present invention is a 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images. After the subject undergoes PC-MRI imaging, the image sequence is input into an image sequence input port, and the image sequence is divided into amplitude and phase image sequences by the image sequence input port.

[0006] The image processing module then generates a maximum velocity image sequence from the amplitude image sequence for vessel segmentation and three-dimensional reconstruction, and automatically measures vessel structure parameters; and measures the time function v(t) of the spatial average blood flow velocity from the phase image sequence.

[0007] The pulmonary artery pressure calculation and display module further extracts three velocity principal components from the time function v(t) of the average flow velocity in the blood flow space, and inputs them together with the vascular structure parameters into the linear regression model to output the average pulmonary artery pressure; at the same time, the pulmonary hypertension diagnosis and evaluation module inputs the three velocity principal components and seven vascular structure parameters into the logistic regression model, outputs the probability of suffering from PH, and determines whether pulmonary hypertension exists.

[0008] The advantages of the present invention are:

[0009] 1. The 4D Flow pulmonary artery pressure measurement and evaluation system of the present invention proposes a method for automatically measuring vascular structure information and blood flow velocity information related to PH, and uses a linear regression model to comprehensively utilize this information.

[0010] 2. The 4D Flow pulmonary artery pressure measurement and evaluation system of the present invention uses a logistic regression model to achieve non-invasive diagnosis of PH.

[0011] 3. The 4D Flow pulmonary artery pressure measurement and evaluation system of the present invention can measure the pulmonary artery in a non-invasive, simple and automated manner, can be repeated multiple times, is convenient for long-term monitoring, and can provide a technical basis for the prognosis and dynamic evaluation of PH. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is an overall schematic diagram of the 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images of the present invention;

[0013] Figure 2 A schematic diagram of the working process of the image processing module of the measurement and evaluation system of the present invention;

[0014] Figure 3 It is a schematic diagram of the principle of generating a maximum speed image sequence from an amplitude image sequence in the measurement and evaluation system of the present invention;

[0015] Figure 4 A schematic diagram of performing blood vessel segmentation from a maximum speed image sequence in the measurement and evaluation system of the present invention;

[0016] Figure 5 A schematic diagram of automatic measurement of three-dimensional blood vessel diameter in the measurement and evaluation system of the present invention;

[0017] Figure 6 It is a schematic diagram of extracting three principal components from the time function v(t) of the spatial average flow velocity of the blood flow in the main pulmonary artery in the measurement and evaluation system of the present invention.

[0018] Figure 7 It is a comparison between the mean arterial pressure measured by the pulmonary artery pressure calculation and display module in the measurement and evaluation system of the present invention and the true value of the catheter pressure measurement. DETAILED DESCRIPTION

[0019] The invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0020] The 4D Flow pulmonary artery pressure measurement and evaluation system of the present invention is as follows: Figure 1 As shown, after the subject undergoes PC-MRI imaging, the image sequence is input into the image sequence input port. The image sequence input port divides the PC-MRI image sequence into an amplitude image sequence and a phase image sequence. The image processing module generates a maximum velocity image sequence from the amplitude image sequence for vascular segmentation and three-dimensional reconstruction, and automatically measures seven vascular structural parameters; at the same time, the time function v(t) of the average blood flow space velocity is measured from the phase image sequence. The pulmonary artery pressure calculation and display module extracts three velocity principal components from the time function v(t) of the average blood flow space velocity, and inputs them into the linear regression model together with the seven vascular structural parameters, outputs the average pulmonary artery pressure, and displays it on the system panel. The pulmonary hypertension diagnosis and evaluation module inputs the three velocity principal components and seven vascular structural parameters into the logistic regression model, and outputs the probability of suffering from PH; at the same time, the system can record patient information and dynamically evaluate the patient's PH development process based on multiple measurement results of the same patient.

[0021] The image sequence in the above-mentioned input image sequence input port is required to be a PC-MRI image sequence in DICOM format; and the image sequence input port judges the image information in the image sequence to check whether it meets the requirements of importing the image sequence input port. The specific judgment method is: if the image spatial resolution is less than 200*200, the layer thickness is greater than 4mm, or the number of phases collected in one cardiac cycle is less than 10, it means that the resolution of the input image sequence does not meet the requirements.

[0022] In this embodiment, the image spatial resolution is 224*224, the layer thickness is 4mm, and the number of acquisition phases in one cardiac cycle is 20, which all meet the requirements of the input image sequence of this system, and thus can be smoothly input into the image sequence input port of this system. For example, if the image information in the input image sequence is: the image spatial resolution is 224*224, the layer thickness is 4mm, and the number of acquisition phases in one cardiac cycle is 20, then it meets the requirements of the input image sequence and can be input into the image sequence input port. Otherwise, the system prompts that the input image sequence does not meet the requirements.

[0023] Then the image sequence input port classifies the received image sequence and reads the IMAGE TYPE field information in the DICOM file information. If it is VELOCITY MAP, it means it is a phase image sequence, which contains the velocity of tissue movement in the three directions of left and right, front and back, and up and down, that is, the velocity vector; if it is M_PCA, it means it is an amplitude image sequence, and the pixel brightness in the amplitude image sequence reflects the amplitude of the movement velocity of the tissue in the image. If it is other, it is other images and will not be processed.

[0024] Finally, the image sequence input port further verifies the classified image sequence, and only imports the image sequence with qualified verification results into the image processing module; the verification qualified conditions are as follows, and the verification is considered qualified only when both verification qualified conditions are met at the same time:

[0025] Condition 1: Read the ROWS and COLUMNS field information of the DICOM file information in the two image sequences, and the values ​​of the two image sequences are equal;

[0026] Condition 2: Read the SLICE LOCATION field information of the DICOM file information in the two image sequences. For any image in the amplitude image sequence, if the read value is h1, there are 3 images in the phase image sequence whose corresponding SLICE LOCATION field information is h1.

[0027] If any of the above conditions is not met, the verification fails, prompting that the image sequence is missing an image and does not meet the requirements; otherwise, it prompts that the image import is successful.

[0028] The above amplitude image sequence and phase image sequence meet the requirements, which is manifested as 800 velocity amplitude images with the IMAGETYPE field of M_PCA and 2400 phase images with VELOCITY MAP in the DICOM file information, satisfying the 1:3 relationship, and the ROW and COLUMNS of the two images are both 224, so the system verification passes.

[0029] By the above method, the amplitude image sequence and the phase image sequence are input into the image processing module, and the image processing module processes them respectively, such as Figure 2 As shown, the processing method is:

[0030] Amplitude image sequence processing:

[0031] First, the pulmonary artery and other blood vessel images around the pulmonary artery, including the ascending aorta, descending aorta, and superior vena cava, are segmented from the amplitude image sequence, and the structural information of the above blood vessels is extracted, including the radius of the main pulmonary artery, the radius of the ascending aorta, the radius of the superior vena cava, and the radius of the descending aorta, and the derived parameters of the ratio of the main pulmonary artery radius to the aorta radius, the ratio of the main pulmonary artery radius to the descending aorta radius, and the ratio of the main pulmonary artery radius to the superior vena cava radius are calculated; the specific method is as follows:

[0032] The maximum brightness value m(x,y) of each pixel (x,y) in all phases of the image is obtained to form a new image sequence Im, in which the pixel brightness represents the maximum speed of tissue movement in a cardiac cycle, so Im is called the maximum speed image sequence, such as Figure 3 As shown in the figure, in a cardiac cycle, the maximum speed of tissue movement can reflect different tissue properties (for example, blood moves faster and the image is brighter, while bones are basically fixed and the image is very dark); using Im can clearly distinguish blood from other tissues, which is convenient for subsequent vascular image segmentation.

[0033] Subsequently, the vascular image is segmented, and grayscale equalization and edge sharpening are performed on Im; since the maximum blood flow velocity in a cardiac cycle is significantly greater than the movement speed of other tissues, the threshold method is used to segment the vascular tissue in the image, and tissues with brightness greater than the threshold T can be segmented as vascular tissue; the segmentation result of the threshold method may contain some holes, which can be eliminated using a closing operation. The above threshold method is used to segment vascular images, and according to the location of the blood vessels, they can be divided into the main pulmonary artery, ascending aorta, descending aorta, and superior vena cava.

[0034] In order to solve the problem of poor blood vessel continuity in the threshold segmentation method, a three-dimensional region growing algorithm is used here to further segment the main pulmonary artery, ascending aorta, descending aorta and superior vena cava into separate, regionally connected, and smooth-edged blood vessel regions, so that the blood vessel continuity is improved, the internal holes of the image are filled, and the interference of other tissues is avoided; Figure 4 As shown, the segmentation result of the main pulmonary artery is shown.

[0035] Finally, the three-dimensional reconstruction of the blood vessels is performed, and the images of each layer of the four blood vessels, the main pulmonary artery, the ascending aorta, the descending aorta, and the superior vena cava, in Im are reconstructed in three dimensions to obtain the three-dimensional models of the four blood vessels. The vascular radius is measured in the three-dimensional model of the four blood vessels, and the radius of the main pulmonary artery, the radius of the ascending aorta, the radius of the superior vena cava, and the radius of the descending aorta are measured. The ratio of the radius of the main pulmonary artery to the radius of the aorta, the ratio of the radius of the main pulmonary artery to the radius of the descending aorta, and the ratio of the radius of the main pulmonary artery to the radius of the superior vena cava are further obtained. So far, a total of seven vascular structure parameters have been obtained. The specific method is:

[0036] Select the image with the largest blood vessel area in the segmented blood vessel image, set the center of gravity of the blood vessel area as the initial point, and 85% of the radius of the inscribed circle of the blood vessel area in the image as the initial blood vessel radius. Use the following algorithm to automatically measure the radius of the three-dimensional blood vessel, such as Figure 5 The algorithm is as follows:

[0037] First, a three-dimensional sphere is generated at the initial point with the initial blood vessel radius as the sphere radius.

[0038] Then, it is determined whether the sphere exceeds the blood vessel area. If not, the radius of the sphere is increased; otherwise, the sphere is moved, and the moving direction and distance are determined by the relative position of the volume of the exceeded part and the center of the sphere. After several iterations, until the sphere can no longer be enlarged and moved, the sphere is the maximum inscribed sphere, and its radius is the radius of the three-dimensional blood vessel.

[0039] The traditional method uses a two-dimensional vessel radius, which is usually not the true vessel radius, but the projection of the three-dimensional vessel radius in the measurement plane. It is affected by the measurement plane and is related to the vessel direction and the experience of the measurement personnel. The three-dimensional vessel radius used in this method avoids manual measurement, reduces human measurement errors, reduces the workload required for clinical examinations, and speeds up the measurement.

[0040] Phase image sequence processing:

[0041] The main pulmonary artery vascular area is determined by segmenting the pulmonary artery image according to the phase image of each phase, and the spatial average blood flow velocity v in the main pulmonary artery is further extracted, and finally the time function v(t) of the spatial average blood flow velocity is formed. The specific method is as follows:

[0042] First, there is a linear relationship between tissue movement velocity V and image brightness I: V = S*I+P, where S and P represent the slope and intercept of velocity coding, respectively, which can be read and obtained in DICOM file information. In this embodiment, the linear relationship between tissue movement velocity V and image brightness I is V = -150*I+0.07326. From this, the blood flow velocity V in the left-right, front-back, and top-bottom directions can be calculated. x , V y , V z .

[0043] Then, the tissue velocity at all points in the main pulmonary artery was integrated to obtain the average velocity in three directions.

[0044]

[0045]

[0046]

[0047] In the above formula, σ represents any point in the main pulmonary artery.

[0048] Finally, the spatial average blood flow velocity v in the main pulmonary artery in each phase is calculated, and the average flow velocity in three directions is obtained:

[0049]

[0050] This constitutes the time function v(t) of the average blood flow velocity in space. For example, if there are 20 imaging phases in the cardiac cycle, v(t) is represented as a vector with a dimension of 20.

[0051] The seven vascular parameters and the time function v(t) of the average blood flow velocity in the above process are sent to the pulmonary artery pressure calculation and display module to calculate the pulmonary artery pressure. The specific method is as follows:

[0052] A. Reduce the dimension of the time function v(t) of the average blood flow velocity in space and extract the first three velocity principal components v1, v2, and v3:

[0053] Assuming that there are T phases in a cardiac cycle, we can use a T-dimensional column vector V T Represents v(t), and its principal component calculation method is:

[0054]

[0055] In the above formula, W T3 is a coefficient matrix with T rows and 3 columns.

[0056] The blood flow information contained in v(t) is thus simplified into three blood flow characteristic values ​​v1, v2, and v3, which represent the average blood flow velocity at the beginning of the systolic period, the peak of the systolic period, and the diastolic period, respectively. There are significant differences between the three blood flow characteristic values ​​in healthy people and PH patients, such as Figure 6 As shown, the three colors represent the blood flow phases corresponding to the three blood flow characteristic values.

[0057] Input v1, v2, v3 and seven vascular structural parameters into the linear regression model, output the mean pulmonary artery pressure p in millimeters of mercury (mmHg), and display it on the system panel. :

[0058]

[0059] In the formula, s i is the i-th vascular structure parameter, where s1~s7 are the radius of the main pulmonary artery, the radius of the ascending aorta, the radius of the descending aorta, the radius of the superior vena cava, the radius of the main pulmonary artery / the radius of the ascending aorta, the radius of the main pulmonary artery / the radius of the descending aorta, and the radius of the main pulmonary artery / the radius of the superior vena cava, respectively; v j is the jth blood flow characteristic value; a i , b j and c are constants, and the values ​​of a1~a7, b1~b3, and c can be obtained by fitting the linear regression model. For example, the mean pulmonary artery pressure of 18 PH patients was measured based on clinical right cardiac catheter pressure measurement, and three blood flow characteristic values ​​were extracted and fitted with seven vascular structure parameters to obtain the parameter a i , b j and c. by a i , b j The mean pulmonary artery pressure calculated by and c is compared with the true value measured by the pressure measuring catheter. Figure 7 , the error is about 7.07%.

[0060] At the same time, the seven vascular parameters and the time function v(t) of the average flow velocity in the blood flow space are also input into the logistic regression model in the pulmonary hypertension diagnosis and evaluation module, and a probability value is output to judge whether pulmonary hypertension exists. The logistic regression model expression is as follows:

[0061]

[0062] Among them, P H is the probability of pulmonary hypertension, α i , β j and γ are all constants, and the values ​​of α1~α7, β1~β3, and γ are obtained by fitting the logistic regression model. H If the probability of pulmonary hypertension is close to 1, it means that the probability of pulmonary hypertension is high; if the probability of pulmonary hypertension isH If the value is close to 0, it means that the probability of suffering from pulmonary hypertension is low. In the present invention, the threshold value 0.5 is set to judge whether the patient suffers from pulmonary hypertension, that is, P H When the value is >0.5, it is diagnosed as pulmonary hypertension. Compared with the current clinical PH diagnostic standard of right heart catheter manometry, the above method has a diagnostic accuracy rate of 100% for PH.

[0063] The above-mentioned pulmonary hypertension diagnosis and evaluation module can also record the measurement information of the subjects, and perform multiple pulmonary artery measurements on the same patient to determine the patient's disease progression and determine whether the patient is in the remission stage or the deterioration stage. Usually, the measurement is performed 1 to 2 times a week. If the average pulmonary artery pressure measured in a single time is greater than 25 mmHg, the system prompts that the pulmonary artery pressure is abnormally elevated; if the average pulmonary artery pressure measured for more than 2 consecutive times is greater than 25 mmHg and shows a gradual increasing trend, the system prompts the risk of acute pulmonary hypertension; if the average pulmonary artery pressure measured for more than 2 consecutive times gradually decreases, the system prompts that the pulmonary artery pressure decreases, indicating that pulmonary hypertension is in the improvement stage. Through multiple measurements of the average pulmonary artery pressure, a dynamic evaluation of pulmonary hypertension can be achieved, thereby evaluating the treatment effect of pulmonary hypertension or warning of the onset of acute pulmonary hypertension.

Claims

1. 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance images, characterized by: After the subject undergoes PC-MRI imaging, the image sequence is input into the image sequence input port, and the image sequence input port divides the image sequence into amplitude and phase image sequences; Then, the image processing module generates a maximum velocity image sequence from the amplitude image sequence for vascular segmentation and three-dimensional reconstruction, and automatically measures the vascular structure parameters; and measures the time function v(t) of the spatial average velocity of blood flow from the phase image sequence; the specific method is: Amplitude image sequence processing: First, the pulmonary artery and other blood vessel images around the pulmonary artery, including the ascending aorta, descending aorta, and superior vena cava, are segmented from the amplitude image sequence, and the structural information of the above blood vessels is extracted, including the radius of the main pulmonary artery, the radius of the ascending aorta, the radius of the superior vena cava, and the radius of the descending aorta, and the derived parameters of the ratio of the main pulmonary artery radius to the aorta radius, the ratio of the main pulmonary artery radius to the descending aorta radius, and the ratio of the main pulmonary artery radius to the superior vena cava radius are calculated; the specific method is as follows: Obtain the maximum brightness value m(x,y) of each pixel point (x,y) in all phases in the image to form a new image sequence Im, where the pixel brightness represents the maximum speed of tissue movement in a cardiac cycle, so Im is the maximum speed image sequence; Then, the vascular image is segmented, and grayscale equalization and edge sharpening are performed on Im; the threshold method is used to segment the vascular tissue in the image, and tissue with brightness greater than the threshold T can be segmented as vascular tissue; the vascular image is segmented by the above threshold method, and the vascular image is divided into the main pulmonary artery, ascending aorta, descending aorta, and superior vena cava according to the location of the blood vessels; Finally, the three-dimensional reconstruction of the blood vessels is performed. The images of each layer of the four blood vessels, namely the main pulmonary artery, ascending aorta, descending aorta and superior vena cava, are reconstructed in three dimensions to obtain the three-dimensional models of the four blood vessels. The vascular radius is measured in the three-dimensional models of the four blood vessels, and the radius of the main pulmonary artery, the radius of the ascending aorta, the radius of the superior vena cava and the radius of the descending aorta are measured. The ratio of the radius of the main pulmonary artery to the radius of the aorta, the ratio of the radius of the main pulmonary artery to the radius of the descending aorta, and the ratio of the radius of the main pulmonary artery to the radius of the superior vena cava are further obtained. So far, a total of seven vascular structure parameters are obtained. The specific method is as follows: Select the image with the largest blood vessel area in the segmented blood vessel images, set the center of gravity of the blood vessel area as the initial point, and 85% of the radius of the inscribed circle of the blood vessel area in the image as the initial blood vessel radius. Use the following algorithm to automatically measure the radius of the three-dimensional blood vessel: First, a three-dimensional sphere is generated at the initial point with the initial blood vessel radius as the sphere radius; Then, it is determined whether the sphere exceeds the blood vessel area. If not, the radius of the sphere is increased; otherwise, the sphere is moved, and the moving direction and distance are determined by the relative position of the volume of the exceeding part and the center of the sphere. After several iterations, the sphere cannot be further increased or moved. At this time, the sphere is the maximum inscribed sphere, and its radius is the radius of the three-dimensional blood vessel. Phase image sequence processing: The main pulmonary artery vascular area is determined by segmenting the phase image of each phase according to the pulmonary artery image, and the spatial average blood flow velocity v in the main pulmonary artery is further extracted, and finally the time function v(t) of the spatial average blood flow velocity is formed. The specific method is as follows: First, there is a linear relationship between tissue movement velocity V and image brightness I: V = S*I+P, where S and P represent the slope and intercept of velocity encoding, respectively, which are read from the DICOM file information; Then, the tissue velocity at all points in the main pulmonary artery was integrated to obtain the average velocity in three directions. In the above formula, σ represents any point in the main pulmonary artery; Finally, the spatial average blood flow velocity v in the main pulmonary artery in each phase is calculated, and the average flow velocity in three directions is obtained: This constitutes the time function v(t) of the spatial average blood flow velocity; Furthermore, the pulmonary artery pressure calculation and display module extracts three velocity principal components from the time function v(t) of the average blood flow velocity in space. Specifically, the feature of the time function v(t) of the average blood flow velocity in space is reduced in dimension to extract the first three velocity principal components v1, v2, and v3: Assuming that there are T phases in a cardiac cycle, a T-dimensional column vector V T Represents v(t), and its principal component calculation method is: In the above formula, W T3 is a coefficient matrix with T rows and 3 columns; Thus, the blood flow information contained in v(t) is simplified into three blood flow characteristic values ​​v1, v2, and v3, which represent the average blood flow velocity at the beginning of the systolic period, the peak of the systolic period, and the diastolic period, respectively; Input v1, v2, v3 and seven vascular structure parameters into the linear regression model and output the mean pulmonary artery pressure p: In the formula, s i is the i-th vascular structure parameter, where s1~s7 are the radius of the main pulmonary artery, the radius of the ascending aorta, the radius of the descending aorta, the radius of the superior vena cava, the radius of the main pulmonary artery / the radius of the ascending aorta, the radius of the main pulmonary artery / the radius of the descending aorta, and the radius of the main pulmonary artery / the radius of the superior vena cava, respectively; v j is the jth blood flow characteristic value; a i , b j and c are all constants, and the values ​​of a1~a7, b1~b3, and c can be obtained by fitting the linear regression model; At the same time, the seven vascular parameters and the time function v(t) of the average flow velocity in the blood flow space are also input into the logistic regression model in the pulmonary hypertension diagnosis and evaluation module, and a probability value is output to judge whether there is pulmonary hypertension; the logistic regression model expression is as follows: Among them, P H is the probability of pulmonary hypertension, α i , β j and γ are all constants, and the values ​​of α1~α7, β1~β3, and γ are obtained by fitting the logistic regression model.

2. The 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance imaging according to claim 1, characterized in that: Images that meet the requirements of image spatial resolution of not less than 200*200, layer thickness of not more than 4mm, or the number of phases acquired in one cardiac cycle of not less than 10 are allowed to be imported into the image sequence port.

3. The 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance imaging according to claim 1, characterized in that: The image sequence input port reads the IMAGE TYPE field information in the DICOM file information. If it is VELOCITYMAP, it means a phase image; if it is M_PCA, it means a velocity amplitude image.

4. The 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance imaging according to claim 1, characterized in that: Image sequences that meet the following conditions are allowed to be imported into the image processing module: Condition 1: The ROWS and COLUMNS fields of the DICOM file information in the amplitude and phase image sequences have the same values ​​for the two image sequences; Condition 2: SLICE LOCATION field information of DICOM file information in the amplitude and phase image sequences. For any image in the amplitude image sequence, if the read value is h1, there are 3 images in the phase image sequence whose corresponding SLICE LOCATION field information is h1.

5. The 4D Flow pulmonary artery pressure measurement and evaluation system based on phase contrast magnetic resonance imaging according to claim 1, characterized in that: The progression of the patient's disease can be judged by performing multiple pulmonary artery measurements on the patient; if the mean pulmonary artery pressure obtained from a single measurement is greater than 25 mmHg, it means that the pulmonary artery pressure is abnormally elevated; if the mean pulmonary artery pressure measured for more than two consecutive times is greater than 25 mmHg and shows a gradually increasing trend, it indicates that there is a risk of acute pulmonary hypertension; if the mean pulmonary artery pressure measured for more than two consecutive times gradually decreases, it means that pulmonary hypertension is in the improvement stage.

Images