Lipid t-lcbi quantification method and system based on multi-modal spatial fusion

Abstract

The application discloses a kind of based on multimodal space fusion's lipid T-LCBI quantification method and system, the method includes obtaining multimodal vascular imaging data registration, realize image distortion correction by cross correlation correction;Based on the boundary of fusion data quantification lipid plaque, based on the fusion data of continuous multiple frames obtains quantified lipid plaque body data;Based on the fusion data of continuous multiple frames constructs blood vessel body data, fuses blood vessel body data with coronary angiography data, based on the spatial trend information of blood vessel calculates the lipid plaque spatial correction coefficient of curved blood vessel section;Based on the quantified lipid plaque body data and spatial correction coefficient, calculate lipid volume;According to the ratio of lipid volume and total voxel volume, calculate T-LCBI index.By registration fusion and distortion correction to multimodal imaging data, break through the bottleneck of lipid boundary quantification, utilize blood vessel body data and coronary angiography data fusion, construct complete spatial trend information, improve the calculation accuracy of LCBI index.

Description

Technical Field

[0001] This invention relates to the field of intravascular imaging technology, and in particular to a lipid T-LCBI quantification method and system based on multimodal spatial fusion. Background Technology

[0002] Cardiovascular imaging technology is mostly developed based on single-modal technology, such as IVUS (intravascular ultrasound), which provides cross-sectional images of blood vessels through high-frequency ultrasound and has made a breakthrough in quantifying plaque burden. However, it cannot distinguish between the lipid core and the fibrous cap microstructure and has low sensitivity in identifying vulnerable plaques. OCT (Optical Coherence Tomography) technology: It uses near-infrared light to achieve micron-level resolution (10-20μm), which can accurately identify fibrous caps, but the penetration depth is only 1.5-2mm, and the trailing edge of deep lipids is not easy to detect.

[0003] NIRS (Near Infrared Spectroscopy) technology generates the LCBI index (lipid core load) through lipid characteristic absorption peaks, but it is limited by two-dimensional integral algorithms, cannot provide depth distribution information, and has low spatial resolution, resulting in a high false detection rate for tortuous blood vessels.

[0004] In view of this, multimodal fusion technology has emerged, such as integrating IVUS / OCT probes to solve the problem of spatiotemporal registration of heterogeneous data. However, due to the differences in the physical principles of OCT and IVUS images, local distortions cannot be mutually corrected. Lipid boundary quantization is still not accurate enough, and single-modal images still have problems such as local distortion. To solve the above problems, this application provides a lipid T-LCBI quantization method and system based on multimodal spatial fusion. Summary of the Invention

[0005] Therefore, the purpose of this invention is to provide a lipid T-LCBI quantification method and system based on multimodal spatial fusion. By registering and correcting anomalies in single-modal imaging data, it breaks through the bottleneck of lipid boundary quantification and uses the fusion of vascular body data and coronary angiography data to calculate vascular spatial orientation information, thereby solving problems such as spatial distortion in three-dimensional reconstruction.

[0006] To achieve the above objectives, this invention provides a lipid T-LCBI quantification method based on multimodal spatial fusion, comprising the following steps: S1. Acquire intravascular OCT images, IVUS images, and coronary angiography images; register the intravascular OCT images and IVUS images; and correct image distortion through cross-correlation correction. S2. The distortion-corrected OCT image and IVUS image are fused together. The boundary of the lipid plaque is quantified in each frame of fused data. The quantified lipid plaque body data is obtained based on the fused data of multiple consecutive frames. S3. Construct vascular body data based on fused data from multiple consecutive frames, fuse vascular body data with coronary angiography data, calculate vascular spatial orientation information based on angiography data, and calculate spatial correction coefficients for lipid plaques in tortuous vascular segments. S4. Calculate lipid volume based on the quantified lipid plaque data from S2 and the spatial correction factor from S3; S5. Calculate the T-LCBI index based on the ratio of lipid volume to total lipid volume.

[0007] More preferably, in step S1, image distortion correction through cross-correlation correction includes: S101. Identify distorted frames in OCT or IVUS images and extract distortion features; S102. Register the distorted frame with a normal image of another modality; S103. The contour is fitted using B-spline curves, and the contour deformation evolution is realized through shape point matching. A rearrangement lookup table is generated to correct the distortion.

[0008] In S103, the vascular contour in the registered OCT image is defined as contour A, and the vascular contour in the registered IVUS image is defined as contour B. Contour A and contour B are represented by B-spline fitting expressions respectively; Based on the matched OCT and IVUS images, all feature points of contours A and B are taken as the shape value points of the B spline contour. The new coordinates of the feature points can be obtained by matching the features of contours B and A. Based on the new coordinate system, control points are derived by inverse calculation, resulting in the undistorted contours of contours A and B after deformation. ; After deformation This generates a rearranged A-line lookup table; by rearranging all A-lines using the rearranged A-line lookup table, the image after distortion correction can be obtained.

[0009] More preferably, the boundary of the quantified lipid plaque in S2 specifically includes: OCT data were used to obtain the position and angular distribution of the lipid front; The location of the posterior lipid margin was determined by combining the medial membrane contour and intensity information from IVUS images. The boundaries of lipid plaques are quantified based on the location of the lipid leading edge, trailing edge, and angular distribution.

[0010] Further preferably, in S3, the vascular body data is fused with coronary angiography data, and the spatial orientation information of the blood vessel is calculated based on the angiography data. The spatial correction coefficient for lipid plaques in tortuous vascular segments is calculated, including: S301. Divide the boundary of the lipid plaque into a first line portion and a second line portion, and calculate the area of ​​the first plaque; S302. Based on coronary angiography data, extract the spatial vessel orientation; calculate the first and second radii of curvature of the vessels in the region where lipid plaques are located; S303. Based on the first radius of curvature and the second radius of curvature, obtain the corresponding center position of the circle; S304. Draw two straight lines from the center of the circle, located on either side of the patch. The angle between the two lines is denoted as the patch angular distribution. Based on the fitting function of the first and second line portions and the patch angle distribution Given the first radius of curvature and the second radius of curvature, calculate the area of ​​the second patch; S305. Determine the spatial correction coefficient based on the ratio of the area of ​​the first patch to the area of ​​the second patch.

[0011] More preferably, in S301, the area of ​​the first patch is calculated using the following formula: in, The area of ​​the first patch. The fitting function for the first line section before correction. The fitting function for the second line section before correction. This represents the height of the plaque.

[0012] More preferably, in S304, the area of ​​the second patch is calculated using the following formula: in, The area of ​​the second patch. This is the fitting function for the corrected first line section. This is the fitting function for the corrected second line section; The first curvature; For the second curvature; This represents the maximum width of the patch, calculated based on NIRS data using OCT.

[0013] Furthermore, in S3, the correction coefficient is calculated using the following formula: in, To fit a function to the first line portion when the data is fused into a long axis view. This is the fitting function for the second line section.

[0014] More preferably, in S4, the lipid volume is calculated based on the corrected lipid plaque, combined with the lipid depth distribution and spatial correction coefficient, using the following formula: in, To identify and quantify lipid depth by combining IVUS and OCT fusion images, , is the calculated depth value for that point. , which represents the number of pixels in the lipid plaques in that region; This is a correction factor; Furthermore, in S5, the T-LCBI index is calculated using the following formula based on the ratio of lipid volume to total lipid volume: in, This is the sum of all pixels involved in the calculation.

[0015] This invention also provides a lipid T-LCBI quantification system based on multimodal spatial fusion for implementing the above method, comprising: The data acquisition module acquires intravascular OCT images, IVUS images, and coronary angiography images, registers intravascular OCT images and IVUS images, and corrects image distortion through cross-correlation correction. The data fusion module fuses the distortion-corrected OCT images and IVUS images, quantifies the boundaries of lipid plaques in each frame of fused data, and obtains quantified lipid plaque body data based on the fused data of multiple consecutive frames. The data correction module constructs vascular body data based on fused data from multiple consecutive frames, fuses the vascular body data with coronary angiography data, calculates the spatial orientation information of the blood vessels based on the angiography data, and calculates the spatial correction coefficient of lipid plaques in tortuous vascular segments. The T-LCBI index calculation module calculates lipid volume based on quantified lipid plaque data and spatial correction coefficients; and calculates the T-LCBI index based on the ratio of lipid volume to total lipid volume.

[0016] The lipid T-LCBI quantification method and system disclosed in this application, based on multimodal spatial fusion, eliminates local distortions caused by uneven duct rotation through OCT / IVUS cross-correlation correction mechanism and B-spline contour evolution and rearrangement lookup table, thereby improving image registration accuracy and solving the spatial distortion problem that cannot be overcome by single-modal correction.

[0017] Accurate calculation of lipid depth is achieved by using the fused IVUS-OCT image and locking the posterior edge of the IVUS middle membrane, breaking through the penetration limit of OCT and the resolution bottleneck of IVUS.

[0018] Based on the radius of curvature extracted from coronary angiography images, spatial correction is performed on lipid plaques in tortuous vessel segments, reducing the detection error of lipid plaques in tortuous vessels.

[0019] This application achieves full integration of "composition-structure-space" elements by integrating the following: data layer: distortion-corrected OCT / IVUS data fusion; feature layer: membrane information and intensity information in IVUS driving NIRS lipid trailing edge correction based on OCT calculation; and spatial layer: coronary angiography providing three-dimensional anatomical orientation, which greatly improves the accuracy of LCBI calculation. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the lipid T-LCBI quantification method based on multimodal spatial fusion of the present invention.

[0021] Figure 2 This is a schematic diagram of the OCT image and its outline according to the present invention.

[0022] Figure 3 This is an IVUS image and its outline schematic diagram of the present invention.

[0023] Figure 4 This invention describes the evolution from distorted contours in OCT images to undistorted contours in IVUS images.

[0024] Figure 5 This is the image rearranged after distortion correction of the contour A of the OCT image according to the present invention.

[0025] Figure 6 This demonstrates the effectiveness of existing technologies in recognizing the trailing edge of lipids in the depth direction.

[0026] Figure 7 This demonstrates the effect of lipid recognition on the fused image in this application.

[0027] Figure 8 This is a schematic diagram illustrating the calculation of spatial correction coefficients after incorporating spatial information into the imaging image in this application.

[0028] Figure 9 This is a schematic diagram of the structure of the lipid T-LCBI quantification system based on multimodal spatial fusion of the present invention. Detailed Implementation

[0029] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] like Figure 1 As shown, one embodiment of the present invention provides a lipid T-LCBI quantification method based on multimodal spatial fusion, comprising the following steps: S1. Acquire intravascular OCT images, IVUS images, and coronary angiography images; register the intravascular OCT images and IVUS images; and correct image distortion through cross-correlation correction. In this process, contrast imaging data and OCT data, or contrast imaging data and IVUS data, are acquired simultaneously. OCT data and IVUS data can be acquired separately or simultaneously. For simultaneous acquisition, the OCT probe and IVUS probe should be in different scanning positions at the same time.

[0031] More preferably, step S1, which achieves image distortion correction through cross-correlation correction, includes: S101. Identify distorted frames in OCT or IVUS images and extract distortion features; S102. Register the distorted frame with a normal image of another modality; S103. The contour is fitted using B-spline curves, and the contour deformation evolution is realized through shape point matching. A rearrangement lookup table is generated to correct the distortion.

[0032] More preferably, the vascular contour in the registered OCT image is defined as contour A, and the vascular contour in the registered IVUS image is defined as contour B. Contour A and contour B are represented by B-spline fitting expressions respectively; Based on the matched OCT and IVUS images, all feature points of contours A and B are taken as the shape value points of the B spline contour. The new coordinates of the feature points can be obtained by matching the features of contours B and A. Based on the new coordinate system, control points are derived by inverse calculation, resulting in the undistorted contours of contours A and B after deformation. ; After shape transformation This generates a rearranged A-line lookup table; by rearranging all A-lines using the rearranged A-line lookup table, the distortion-corrected image can be obtained.

[0033] Distortion occurs due to uneven rotation caused by significant differences in the forces exerted on the catheter. The following example illustrates this phenomenon, where OCT data is distorted, but the corresponding IVUS data remains undistorted, and the IVUS data is used to correct the OCT data.

[0034] Figure 2This is a schematic diagram of the OCT image outline. Z1-Z2-Z3-Z4-Z1 represents the acquired A-line data, and W1-W2-W3-W4-W1 represents the actual physical position of the A-line in the circumferential direction. Looking clockwise, the data from Z1 to Z4 should actually correspond to the physical data from W2 to W3. This is due to distortion, which stretches the imaging data (Z1-Z4, the slowly rotating scan line) from the physical W2 to W3 in the circumferential direction into the W1 to W4 region. To correct the distortion, the data from Z1 to Z4 needs to be compressed into W2-W3. Figure 3 The outline and schematic diagram of IVUS are shown below. Figure 2 The image frames with OCT distortion are matched with the undistorted IVUS image frames. The vessel contour acquired by OCT is denoted by A, and the corresponding vessel contour acquired by IVUS is shown by B. Figure 2 All line data in the dataset is uniformly distributed. To correct distortion, the data needs to be rearranged. The rearranged data is as follows: Figure 4 As shown in the right figure. Based on image registration, contour A and contour B are registered. Based on feature information, feature information points of contour A and contour B are registered. Finally, contour A is deformed to evolve into a normal contour without distortion. Based on the feature point distribution of the normal contour, all A-lines of contour A are rearranged to complete the correction of distortion.

[0035] It should be noted that, Figure 2 The imaging is arranged solely according to the sequence number of the A-line data, with the lines arranged uniformly, without considering whether the data acquisition was achieved through uniform or non-uniform rotation. Figure 4 The uniformity of the line arrangement was adjusted to take into account the difference between distorted and undistorted rotations. Figure 5 This is an image rearranged after distortion correction of distorted frames from OCT.

[0036] The evolution from contour A in the OCT image to contour B in the IVUS image is as follows: First, an expression for B-spline fitting is used to represent the irregular closed curve. The expression for the contour A of the OCT image is as follows: in: here Indicates control points, Represents a node. express Second-rate.

[0037] Here, we choose a cubic B-spline curve, so the contours A and B are represented as follows: Here, we choose the same B-spline basis function, so contours A and B are represented as follows: By matching features from OCT and IVUS images, all feature points of the OCT contour A are used as shape points for the spline contour B. New coordinates of these feature points are obtained by matching the IVUS contour B with these new shape points. Then, control points are calculated based on these new coordinate shape points to obtain the undistorted contour of the OCT contour A after deformation. After shape transformation This generates a rearranged A-line lookup table. By rearranging all A-lines using the rearranged A-line lookup table, the distortion-corrected image can be obtained.

[0038] Similarly, the above method also applies if an IVUS image at a certain location shows distortion while an OCT image does not.

[0039] Data fusion is performed on the corrected images. The fusion method can employ existing technologies, such as methods in the applicant's prior patent applications, including methods for three-dimensional fusion of dual-mode coronary artery images and methods within fusion systems. As mentioned earlier, NIRS lipid data is obtained using OCT. Due to rapid light decay, the recognition effect for the trailing edge of thicker lipid plaques in the depth direction is as follows... Figure 6 As shown, there are inaccuracies. After fusing OCT and IVUS, the high penetration of IVUS, especially its ability to identify middle membrane information, can serve as effective information for determining the posterior margin of lipids. For example... Figure 7 The information is synthesized and corrected. The green curve in the figure represents the middle membrane information, and the yellow curve represents the lipid information determined by a combination of IVUS and NIRS. The angle between the two straight lines corresponds to the surrounding angle of the lipid information in the cross-sectional image.

[0040] S2. The distortion-corrected OCT image and IVUS image are fused together. The boundary of the lipid plaque is quantified in each frame of fused data. The quantified lipid plaque body data is obtained based on the fused data of multiple consecutive frames. More preferably, the boundary of the quantified lipid plaque in S2 specifically includes: OCT images were used to obtain the position and angular distribution of the lipid front. The location of the posterior lipid margin was determined by combining the medial membrane contour and intensity information from IVUS images. The boundaries of lipid plaques are quantified based on the location of the lipid leading edge, trailing edge, and angular distribution.

[0041] Here, IVUS and OCT fused images are introduced, and lipid recognition is performed on the fused images.

[0042] The first step is to identify lipid information based on OCT spectral signals (NIRS; for specific identification methods, please refer to the method in the applicant's prior application, No. 202311584776.8, entitled "Lipid Plaque Detection and Analysis Method and System Based on OCT Imaging"). This can accurately identify the angular distribution of lipids and the C1 distribution at the starting position of lipid depth. In the depth direction, qualitative lipid components can be identified for thicker lipid plaques, but the posterior boundary of lipids cannot be identified.

[0043] The second step, based on the high penetration effect of IVUS on blood vessels, accurately identifies the vascular media and locates the annular contour of the vascular media. Then, combined with the intensity information identified by IVUS, the posterior edge C2 of the lipid plaque is determined.

[0044] Based on steps one and two, the precise angles and front and rear edge distributions C1 and C2 can be calculated, thus obtaining quantified lipid distribution information. For example... Figure 7 Part 1 of the central region allows for precise adjustment of the NIRS lipid distribution ring calculated based on OCT. The adjusted lipid distribution ring is the outer ring shown in the figure, and the lipid region is the area marked in yellow in the figure.

[0045] Therefore, by combining the first and second steps, the location and size of the lipid pool can be precisely determined. It should be noted that "first step" and "second step" are used for ease of description only and do not specify the order of the steps.

[0046] The above processing was performed on all OCT and IVUS image frames to obtain complete lipid plaque data.

[0047] S3. Construct vascular body data based on fused data from multiple consecutive frames, fuse vascular body data with coronary angiography data, calculate vascular spatial orientation information based on angiography data, and calculate spatial correction coefficients for lipid plaques in tortuous vascular segments. Specifically: S301. Divide the boundary of the lipid plaque into a first line part and a second line part; here, the boundary of the lipid plaque is divided into two parts, represented by different lines, and different function expressions are constructed. The division method is to use an axis perpendicular to the major axis as the boundary, which passes through the highest point of the plaque. The function expression of the first line part (the thick red line on the left) is: The function expression for the second line section (the thin black line on the right) is: The height of the plaque is .

[0048] In the long axis mode, the area expression for the first patch is: The area of ​​the first patch. It is calculated from plaque data obtained by fusion of OCT and IVUS data. The long axis mode does not consider the tortuosity of blood vessels. (See reference...) Figure 8 The image above.

[0049] S302. Based on coronary angiography data, the spatial vascular orientation is extracted; the curvature of segmented blood vessels can be obtained, thereby obtaining the first and second radii of curvature of blood vessels in the region where lipid plaques are located.

[0050] The size of the plaque in the direction of the vessel curvature radius is ,satisfy ,and , The first radius of curvature, The second radius of curvature.

[0051] S303. Based on the first radius of curvature and the second radius of curvature, obtain the corresponding center position of the circle; S304. Draw two straight lines from the center of the circle, located on either side of the patch. The angle between the two lines is denoted as the patch angular distribution. Based on the fitting function of the first and second line portions and the patch angle distribution and the first radius of curvature Second radius of curvature Calculate the area of ​​the second plaque, that is, the area of ​​the plaque considering the tortuous state of the blood vessel; More preferably, in S304, the area of ​​the second patch is calculated using the following formula: in, The area of ​​the second patch. This is the fitting function for the corrected first line section. This is the fitting function for the corrected second line section; The first radius of curvature; The second radius of curvature; This represents the maximum width of the patch, calculated based on patch volume data obtained from the fusion of OCT and IVUS data.

[0052] S305, determine the spatial correction coefficient based on the ratio of the area of ​​the first patch to the area of ​​the second patch.

[0053] The spatial correction factor for lipid plaques in tortuous vascular segments is calculated using the following formula: in and To correct the fitting function of the first line section before and after, and This is the fitting function for correcting the second line section before and after.

[0054] S4. Calculate lipid volume based on the quantified lipid plaque data from S2 and the spatial correction factor from S3; Specifically, the lipid volume can be calculated using the following formula: in, To identify and quantify lipid depth by combining IVUS and OCT fusion images, , is the calculated depth value for that point. , which represents the number of pixels in the lipid plaques in that region; This is a correction factor; S5. Calculate the T-LCBI index based on the ratio of lipid volume to total lipid volume.

[0055] In S5, the T-LCBI index is calculated using the following formula: in, This is the sum of all pixels involved in the calculation.

[0056] This invention also provides a lipid T-LCBI quantification system based on multimodal spatial fusion, used to implement the above-mentioned lipid T-LCBI quantification method based on multimodal spatial fusion, comprising: The data acquisition module acquires intravascular OCT images, IVUS images, and coronary angiography images, registers intravascular OCT images and IVUS images, and corrects image distortion through cross-correlation correction. The data fusion module fuses the distortion-corrected OCT images and IVUS images to quantify the boundaries of lipid plaques and obtain quantified lipid plaque body data based on the fused data of multiple consecutive frames. The data correction module constructs vascular body data based on fused data from multiple consecutive frames, fuses the vascular body data with coronary angiography data to calculate the spatial orientation information of the blood vessels, and performs spatial correction on lipid plaques in tortuous vascular segments. The T-LCBI index calculation module calculates lipid volume based on quantified lipid plaque data and spatial correction coefficients; and calculates the T-LCBI index based on the ratio of lipid volume to total lipid volume.

[0057] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A lipid T-LCBI quantification method based on multimodal spatial fusion, characterized in that: S1. Acquire intravascular OCT images, IVUS images, and coronary angiography images; register the intravascular OCT images and IVUS images; and correct image distortion through cross-correlation correction. S2. The distortion-corrected OCT image and IVUS image are fused together. The boundary of the lipid plaque is quantified in each frame of fused data. The quantified lipid plaque body data is obtained based on the fused data of multiple consecutive frames. S3. Construct vascular body data based on fused data from multiple consecutive frames, fuse the vascular body data with coronary angiography data, calculate the spatial orientation information of the blood vessels based on the angiography data, and calculate the spatial correction coefficient for lipid plaques in tortuous vascular segments; specifically including: S301. Divide the boundary of the lipid plaque into a first line portion and a second line portion, and calculate the area of ​​the first plaque; S302. Based on coronary angiography data, extract the spatial vessel orientation; calculate the first and second radii of curvature of the vessels in the region where lipid plaques are located; S303. Based on the first radius of curvature and the second radius of curvature, obtain the corresponding center position of the circle; S304. Draw two straight lines from the center of the circle, located on either side of the patch. The angle between the two lines is denoted as the patch angular distribution. Based on the fitting function of the first and second line portions and the patch angle distribution Given the first radius of curvature and the second radius of curvature, calculate the area of ​​the second patch; S305. Determine the spatial correction coefficient based on the ratio of the area of ​​the first patch to the area of ​​the second patch; S4. Calculate lipid volume based on the quantified lipid plaque data from S2 and the spatial correction factor from S3; S5. Calculate the T-LCBI index based on the ratio of lipid volume to total lipid volume.

2. The lipid T-LCBI quantification method based on multimodal spatial fusion according to claim 1, characterized in that, In step S1, image distortion correction through cross-correlation correction includes: S101. Identify distorted frames in OCT or IVUS images and extract distortion features; S102. Register the distorted frame with a normal image of another modality; S103. The contour is fitted using B-spline curves, and the contour deformation evolution is realized through shape point matching. A rearrangement lookup table is generated to correct the distortion; including: defining the vascular contour in the registered OCT image as contour A, and defining the vascular contour in the registered IVUS image as contour B. Contour A and contour B are represented by B-spline fitting expressions respectively; Based on the matched OCT and IVUS images, all feature points of contours A and B are taken as the shape value points of the B spline contour. The new coordinates of the feature points can be obtained by matching the features of contours B and A. Based on the new coordinate system, control points are derived by inverse calculation, resulting in the undistorted contours of contours A and B after deformation. ; After deformation This generates a rearranged A-line lookup table; by rearranging the A-line lookup table, all A-lines are rearranged to obtain the image after distortion correction.

3. The lipid T-LCBI quantification method based on multimodal spatial fusion according to claim 1, characterized in that, The boundaries of the quantified lipid plaques in S2 include: OCT images were used to obtain the position and angular distribution of the lipid front. The location of the posterior lipid margin was determined by combining the medial membrane contour and intensity information from IVUS images. The boundaries of lipid plaques are quantified based on the location of the lipid leading edge, trailing edge, and angular distribution.

4. The lipid T-LCBI quantification method based on multimodal spatial fusion according to claim 1, characterized in that, The area of ​​the first patch in S301 is calculated using the following formula: in, The area of ​​the first patch. The fitting function for the first line section before correction. The fitting function for the second line section before correction. This represents the height of the plaque.

5. The lipid T-LCBI quantification method based on multimodal spatial fusion according to claim 4, characterized in that, In S304, the area of ​​the second patch is calculated using the following formula: in, The area of ​​the second patch. This is the fitting function for the corrected first line section. This is the fitting function for the corrected second line section; The first curvature; For the second curvature; This indicates the maximum width of the patch.

6. The lipid T-LCBI quantification method based on multimodal spatial fusion according to claim 5, characterized in that, In S305, the correction factor is calculated using the following formula: 。 7. A lipid T-LCBI quantification system based on multimodal spatial fusion, used to implement the method according to any one of claims 1-6, comprising: The data acquisition module acquires intravascular OCT images, IVUS images, and coronary angiography images, registers intravascular OCT images and IVUS images, and corrects image distortion through cross-correlation correction. The data fusion module fuses the distortion-corrected OCT images and IVUS images, quantifies the boundaries of lipid plaques in each frame of fused data, and obtains quantified lipid plaque body data based on the fused data of multiple consecutive frames. The data correction module constructs vascular body data based on fused data from multiple consecutive frames, fuses the vascular body data with coronary angiography data, calculates the spatial orientation information of the blood vessels based on the angiography data, and calculates the spatial correction coefficient of lipid plaques in tortuous vascular segments. The T-LCBI index calculation module calculates lipid volume based on quantified lipid plaque data and spatial correction coefficients; and calculates the T-LCBI index based on the ratio of lipid volume to total lipid volume.

Images