An AI-driven method and system for three-dimensional analysis of left atrial CT and automatic assessment of indentation areas

Abstract

This invention provides an AI-driven method and system for three-dimensional analysis of left atrial CT and automatic assessment of indentation regions. The method includes: acquiring three-dimensional medical image data of the patient's heart; performing structural segmentation of the left atrium and adjacent organs based on a deep learning model; automatically identifying indentation regions using a spatial distance calculation strategy; extracting, simplifying, and smoothing the indentation contour, and calculating its perimeter and area; and finally, visually displaying the morphology and distribution of the indentation regions through two-dimensional overlay and three-dimensional reconstruction. This invention achieves automatic identification, quantitative analysis, and multi-dimensional visualization of indentation regions, which helps improve the efficiency and accuracy of preoperative planning for atrial fibrillation and has good clinical application value.

Description

Technical Field

[0001] This invention belongs to the field of medical image processing technology, specifically relating to a deep learning-based medical image analysis method, particularly for organ segmentation, measurement, and risk assessment of preoperative CT images in patients with atrial fibrillation. This invention provides an automated method for analyzing the heart and its surrounding important anatomical structures to assist clinicians in surgical planning and risk assessment. Background Technology

[0002] Atrial fibrillation is one of the most common sustained arrhythmias, with high morbidity and mortality, and has become a significant public health problem worldwide. Catheter ablation, as one of the main treatments for atrial fibrillation, may cause damage to adjacent important anatomical structures such as the esophagus and coronary sinus during the procedure, leading to serious complications. Therefore, accurate preoperative assessment of the spatial relationship between the left atrium and its surrounding anatomical structures is crucial for improving surgical safety.

[0003] Currently, preoperative assessment mainly relies on doctors manually identifying and measuring organs on CT images. This method has the following problems:

[0004] 1. Manual analysis is time-consuming and labor-intensive, making it difficult to meet clinical needs;

[0005] 2. The analysis results are easily affected by the operator's experience and subjective judgment, making it difficult to guarantee accuracy and consistency;

[0006] 3. It is difficult to perform precise quantitative analysis on complex three-dimensional anatomical structures;

[0007] 4. The lack of standardized procedures and objective evaluation indicators makes it difficult to achieve comparability between clinical settings.

[0008] In recent years, with the development of artificial intelligence technology, especially the breakthrough progress of deep learning in the field of medical image segmentation, new technical means have been provided for the automated analysis of the heart and its surrounding structures. However, existing methods mainly focus on the segmentation of single organs, lacking comprehensive analysis of multiple related anatomical structures, and have not specifically addressed the problems of automatic identification and quantitative assessment of indentation areas.

[0009] Therefore, there is an urgent need for a method that can automatically, accurately, and efficiently analyze preoperative CT images of patients with atrial fibrillation, to locate, quantify, and visualize the spatial indentation between the left atrial wall and adjacent organs, and to provide clinicians with a more comprehensive basis for preoperative planning. Summary of the Invention

[0010] This invention combines artificial intelligence technology and image processing algorithms to provide an AI-driven method for three-dimensional analysis of left atrial CT scans and automatic assessment of indentation areas, including:

[0011] Step 1: Acquire three-dimensional medical image data of the patient's heart;

[0012] Step 2: Based on deep learning models or image segmentation algorithms, extract the relevant structures of the left atrium from the three-dimensional medical image data, including the left atrium, left atrial wall, esophagus, coronary sinus, aorta, left atrial appendage, and pulmonary veins;

[0013] Step 3: Utilizing the spatial relationship between the left atrial wall and adjacent organs, and based on anatomical structure labeling and spatial distance calculation methods, the indentation area is automatically located and point set is extracted;

[0014] Step 4: Extract and quantify the geometric contour of the identified left atrial indentation area;

[0015] Step 5: Visualization of the results of multiple indentations in the left atrium.

[0016] Furthermore, in step 1, the three-dimensional medical image data can be any medical image data containing complete information about the left ventricular structure, including CT images or MRI images.

[0017] Furthermore, in step 2, medical software is used to manually annotate the structures, or artificial intelligence deep learning algorithms are used to automatically segment the main anatomical structures of the left ventricular region. The deep learning algorithms include STU-Net-B and SwinUNETR. After segmentation, the left ventricular region consists of the left atrium, left atrial wall, coronary sinus, aorta, and esophagus.

[0018] Furthermore, step 3 includes the following steps:

[0019] Step 3.1: Based on the anatomical structure segmentation results obtained from the deep learning model, extract the three-dimensional label map of each anatomical structure; the label map is a three-dimensional voxel grid data structure, where each voxel position has a corresponding label value and three-dimensional spatial coordinate index, wherein the left atrial wall, aorta, coronary sinus and esophagus are assigned different label values; in order to improve the accuracy and stability of subsequent calculations, the segmentation results are preprocessed, including removing isolated noise points and filling small holes, so as to ensure the boundary continuity and structural integrity of the label region;

[0020] Step 3.2: Construct a label-mask-based spatial distance calculation strategy; using the label mask of the left atrial wall as the reference region, perform voxel-level spatial distance calculations with the target regions represented by the label masks of the aorta, coronary sinus, and esophagus respectively; for each pair of voxels between the above reference region and the target region, use the fast Euclidean distance transformation algorithm to calculate its shortest spatial distance, the specific calculation method is as follows:

[0021]

[0022] Where p1(x1,y1,z1) and p2(x2,y2,z2) are the spatial coordinates of any two voxels in any two label images; by setting a distance threshold, preferably 2 mm, all voxel point pairs with a distance less than the threshold are selected as the point set of potential indentation regions; to improve the efficiency of calculation and selection, spatial indexing is performed on all label voxels based on the KD tree data structure to accelerate the search process of nearest neighbor voxels;

[0023] Step 3.3: Perform contour extraction and geometric optimization on the acquired indentation area point set; first, project the 3D point set onto a preset 2D plane to obtain the projected 2D point set of the indentation area:

[0024]

[0025] Among them, (x i ,y i () represents the coordinates of the indentation point on a two-dimensional plane;

[0026] Based on this point set, the Alpha-shape algorithm is used to extract the boundary contour of the indentation region. The contour is defined as follows:

[0027] Γ α =Alpha-shape(P indentation ,α)

[0028] Where α is a parameter that controls the complexity of the boundary shape, and its value affects the degree of contour shrinkage and the effect of local feature preservation.

[0029] After obtaining the initial contour, the boundary contour is further geometrically optimized: First, the contour is simplified using the Douglas-Peucker algorithm, removing redundant vertices while maintaining the main geometric features of the contour; then, B-spline curves are applied to smooth the simplified contour to improve its continuity and smoothness, providing optimized boundary input for subsequent quantitative analysis.

[0030] Furthermore, in step 4, the analysis and calculation of the indentation area includes the following steps:

[0031] Step 4.1: Based on the extracted indentation contour data, calculate the perimeter of the indentation area. The contour is typically composed of several straight line segments and curved segments. To calculate the perimeter, the lengths of each line segment must first be summed. For straight line segments, their lengths can be directly calculated based on their endpoint coordinates. For curved segments, a piecewise linear approximation method is used to approximate them into multiple short line segments to improve the accuracy of the calculation. By summing the lengths of all contour line segments, the total perimeter of the indentation area is obtained. The key to this process is to perform accurate approximation calculations on the curved portion to ensure the accuracy of the overall perimeter.

[0032] Step 4.2: Since the outline of the indentation area is a closed curve in three-dimensional space, it can be processed by the area calculation method of three-dimensional polygonal patches. Specifically, the spatial surface enclosed by the outline is triangulated into multiple triangular patches, the area of ​​each triangular patch is calculated, and the areas are summed to obtain the total area of ​​the entire indentation area.

[0033] Furthermore, in step 5, the original CT image is first registered and superimposed with the automatically segmented structural mask to accurately display the boundaries of each organ. Subsequently, the mask is reconstructed in three dimensions using 3D reconstruction technology, so that the indentation area is clearly presented in a structured form in three-dimensional space. Finally, the indentation area, CT image, and reconstructed mask are comprehensively superimposed to intuitively present the spatial location, morphology, and characteristics of the indentation and its relationship with the surrounding anatomical structures, providing a comprehensive assessment of the distribution and risk characteristics of the indentation area from multiple perspectives.

[0034] The present invention also provides an AI-driven three-dimensional analysis and automatic assessment system for indentation areas of left atrial CT for implementing the above method, the system comprising:

[0035] The medical image acquisition module is used to acquire three-dimensional medical image data containing the left atrial region;

[0036] The structural segmentation module is used to identify anatomical structures in the image data and extract the left atrium, left atrial wall, esophagus, aorta and coronary sinus, wherein the structural segmentation is achieved by an artificial intelligence deep learning model or a medical image processing algorithm.

[0037] The indentation localization module is used to spatially locate and extract point sets of the indentation area between the left atrial wall and adjacent organs based on the label map and spatial distance calculation strategy.

[0038] The geometric calculation module is used to extract the contour, calculate the perimeter and calculate the area of ​​the indentation area. The contour extraction includes Alpha-shape contour construction, Douglas-Peucker algorithm simplification and B-spline curve smoothing. The area calculation includes dividing the projected surface of the indentation area into multiple triangular patches and calculating the area of ​​each patch.

[0039] The visualization module is used to fuse the indentation area with CT images and structural masks to present the spatial location, geometric shape, and relative relationship of the indentation to the anatomical structure, which is used to assist in preoperative clinical assessment.

[0040] Technical effect

[0041] This invention utilizes an AI-driven method for preoperative CT 3D analysis and automated assessment of indentation areas in atrial fibrillation (AF) surgery, achieving precise analysis of CT images and automated assessment of indentation regions. This method accurately identifies the heart and surrounding structures through automatic segmentation to generate masks and perform 3D reconstruction, enabling precise localization and morphological analysis of the indentation region. Compared to traditional manual analysis, this method significantly improves the accuracy and consistency of assessment, avoids interference from human factors, and significantly shortens analysis time. Furthermore, the visualization of the indentation region enhances the understanding of indentation morphology and distribution, provides clinicians with an intuitive reference, improves the accuracy of preoperative AF assessment, and strongly supports the development of personalized treatment plans. Attached Figure Description

[0042] Figure 1 A flowchart of an AI-driven method for three-dimensional analysis of left atrial CT and automatic assessment of indentation areas;

[0043] Figure 2 This is a schematic diagram showing the segmentation of key anatomical structures such as the left atrial wall, esophagus, and aorta described in step 2.

[0044] Figure 3 This is a schematic diagram of the spatial positioning process of the indentation area described in step 3;

[0045] Figure 4 This is a schematic diagram of the geometric measurement results of the indentation area described in step 4;

[0046] Figure 5 This is a visualization of the indentation area under the three-dimensional fusion view described in step 5. Detailed Implementation

[0047] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that the embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Furthermore, it should be understood that after reading the disclosure of this invention, those skilled in the art can make various modifications or alterations to the invention, and these equivalent forms also fall within the scope of protection defined by this invention.

[0048] This invention proposes an AI-driven automatic assessment method for the indentation region of the left atrium and adjacent organs based on medical 3D imaging data, such as... Figure 1 As shown, it includes the following steps:

[0049] Step 1: Acquire 3D medical image data of the patient's heart. This step is mainly used to acquire preoperative cardiac CT image data for patients with atrial fibrillation. Specifically, during the acquisition process, it is necessary to ensure that the image data completely covers the left atrium and its adjacent key anatomical structures, including but not limited to the aorta, coronary sinus, esophagus, and other important organs that may leave indentations with the left atrium. Furthermore, the acquired medical image data needs to undergo standardized preprocessing, including image quality optimization and spatial registration, to ensure the accuracy of subsequent indentation analysis. The preprocessing includes image noise suppression, contrast enhancement, voxel size standardization, and spatial registration with a standard anatomical template. The preprocessed image data should meet the spatial resolution and sharpness standards required for preoperative planning, providing stable input for subsequent anatomical structure segmentation and automatic identification of indentation areas.

[0050] Step 2: Extracting relevant anatomical structures of the left atrium from 3D medical image data. This step addresses the preoperative indentation assessment needs of atrial fibrillation surgery, focusing on extracting key anatomical structures in the left atrium and its adjacent areas that may form indentations. Special attention is paid to the spatial relationships of structures such as the left atrial wall, aorta, coronary sinus, and esophagus. The segmentation results are shown below. Figure 2 As shown.

[0051] In this step, image segmentation can be achieved through manual annotation or automated segmentation based on artificial intelligence. For manual annotation, medical image processing software such as 3D Slicer or MIMICS can be used to have doctors perform 3D annotation of key structures. For automated segmentation, deep learning-based medical image segmentation network models, such as STU-Net-B (Scalable and Transferable Medical Image Segmentation Models Empowered by Large-Scale Supervised Pre-training) or SwinUNETR (Swin UNEt TRansformers), can be used to segment CT images structurally.

[0052] The structural segmentation results generated using any of the above methods will be converted into structural label maps, representing the spatial distribution of each anatomical structure at the voxel level. To ensure the accuracy of the label maps, the segmentation results are further processed morphologically, including removing isolated voxels and filling small gaps, thereby improving the continuity of label boundaries and anatomical consistency.

[0053] Step 3: Spatial localization of indentation areas. This step, based on the segmentation results obtained in Step 2, achieves precise localization of indentation areas that may form between the anatomical structures of adjacent organs in the left atrium. The identification results are as follows: Figure 3 As shown, the specific steps include the following:

[0054] Step 3.1: Extracting Anatomical Structure Labels. Based on the segmentation results of the deep learning model, labels are extracted for key anatomical structures such as the left atrial wall, aorta, coronary sinus, and esophagus. Specifically, a unique label value is assigned to each anatomical structure, and a standardized three-dimensional spatial coordinate index system is constructed to support subsequent distance calculations and spatial registration operations. Furthermore, to ensure the structural continuity and spatial integrity of the label images, morphological optimization processing is performed on the extracted label images, including removing isolated noise points (such as isolated small regions) formed during segmentation, small holes in the label images, and repairing discontinuities in structural boundaries.

[0055] Let the original set of label images be:

[0056]

[0057] The optimized label image is represented as follows:

[0058]

[0059] L i This represents the label region for the i-th anatomical structure. This indicates the operation of removing noise from the label graph. The process of filling the holes inside the label area. It is an optimized label image with a complete structure and continuous boundaries after removing noise and filling holes.

[0060] Step 3.2: Construct a spatial distance calculation strategy based on the label map. Specifically, based on the label mask region of the left atrial wall, calculate the shortest distance between it and the label mask regions of the aorta, coronary sinus, and esophagus to identify areas where anatomical compression or contact may exist. Further, use the fast Euclidean distance transform algorithm to calculate the minimum three-dimensional distance between any voxel pair between two label regions, thus realizing spatial distance calculation. The specific calculation formula is as follows:

[0061]

[0062] Where p1(x1,y1,z1) and p2(x2,y2,z2) represent the spatial coordinates of the left atrial wall relative to any two voxels in any two label images of the target organ, respectively, and d euclidean It represents the spatial distance between two points.

[0063] Furthermore, combining clinical experience and anatomical tolerance during atrial fibrillation surgery, a distance threshold is set, preferably 2 mm, and all voxel point pairs with a distance less than this threshold are selected as the point set of potential indentation regions. This distance calculation strategy fully reflects the proximity relationships between anatomical structures and can serve as the basic data support for subsequent indentation region extraction and contour recognition.

[0064] Step 3.3: Indentation Contour Extraction and Optimization. This step extracts the contour and optimizes the geometric shape of the indentation area point set obtained in Step 3.2. Specifically, it includes the following three stages:

[0065] First, the three-dimensional indentation point set is projected onto a preset two-dimensional observation plane to obtain the two-dimensional point set of the indentation area, denoted as:

[0066]

[0067] Among them, (x i ,y i ) is the set of indentation points projected onto a two-dimensional plane.

[0068] Applying the Alpha-shape algorithm to the point set P indentation Contour construction is performed to extract the initial boundary contour of the point set:

[0069]

[0070] Here, α is a key parameter of the Alpha-shape algorithm, which is used to control the shape complexity. The smaller α is, the tighter the outline, which is suitable for preserving local structural features.

[0071] Subsequently, the Douglas-Peucker algorithm was used to refine the initial contour. Simplify the process by removing redundant vertices while maintaining the overall geometry:

[0072]

[0073] Where ∈ is the simplification error tolerance, used to control the maximum allowable contour deviation during the simplification process, C simp This represents the simplified outline.

[0074] Finally, B-spline curves (B-Spline) are used to smooth the simplified contour, improving its continuity and smoothness, and generating the final optimized contour curve.

[0075]

[0076] Where k is the order of the B-spline curve, and C final P represents indentation The final optimized contour obtained after the point set is processed by Alpha-shape, DouglasPeucker, and B-Spline algorithms can be directly used for subsequent geometric quantitative analysis such as perimeter and area, and supports 3D spatial reconstruction and visualization rendering.

[0077] Step 4: Quantitative Analysis and Calculation of Indentation Area. Based on the spatial localization and contour extraction of the indentation area, this step systematically quantifies its geometric features, outputting key parameters that can be used to assess the degree of indentation. The results are as follows: Figure 4 As shown, the specific steps include the following:

[0078] Step 4.1: Calculation of the perimeter of the indentation area. This step is used to calculate the accurate perimeter of the indentation area contour. Specifically, the indentation area contour is segmented, and the straight segments and curved segments are processed separately. For straight segments, the length is directly calculated using the distance formula between two points in space; for curved segments, a piecewise linear approximation method is used to approximate them into several short straight segments, and the lengths of each segment are calculated and accumulated. Let the contour consist of n sequentially arranged points, each point being:

[0079] p i =(x i ,y i ,z i ), i = 1, 2, ..., n

[0080] The formula for calculating the total perimeter L is:

[0081]

[0082] Among them, ||p i+1 -p i || represents the spatial distance of the i-th segment. This parameter is used to measure the overall length of the indentation region boundary, providing an important basis for indentation range assessment.

[0083] Step 4.2: Calculation of Indentation Area. This step calculates the spatial area of ​​the indentation region to quantify its severity. Specifically, considering the complexity of the indentation region in three-dimensional space, the area calculation method using three-dimensional polygonal patches is adopted. Further, the indentation region is divided into multiple triangular patches through triangulation, and the area of ​​each patch is calculated and summed to obtain the total area of ​​the indentation region.

[0084] Suppose the contour consists of n points arranged in a sequential order, each point being:

[0085] p i =(x i ,y i ,z i ), i = 1, 2, ..., n

[0086] The center point of the contour is:

[0087]

[0088] By p center With two consecutive contour points p i p i+1 If the i-th triangle can be formed, then the three vertices of the i-th triangle are:

[0089] The area of ​​the triangle is:

[0090]

[0091] The total area A of the indentation region is the sum of the areas of n-1 triangles:

[0092]

[0093] This area parameter, as one of the core indicators of indentation quantification results, can be used to support clinical analysis of indentation depth and the risk to adjacent tissues.

[0094] Step 5: Visualize the indentation analysis results, as shown below. Figure 5 As shown. This step uses a 2D and 3D fusion method to visually present the identification results of the indentation area, facilitating preoperative assessment and multi-angle observation by the doctor. It mainly includes the following two sub-steps:

[0095] Step 5.1: Two-dimensional section display. In this sub-step, the automatically segmented structural mask is precisely registered with the original CT image and overlaid on standard anatomical sections (such as coronal, sagittal, and axial planes). Different colors are used to mark the boundaries of each structure and the indentation area, clearly showing the spatial relationship between the boundaries of each anatomical structure and the indentation area.

[0096] Step 5.2: 3D Reconstruction and Display. This step displays the indentation analysis results in three-dimensional space. Specifically, 3D reconstruction is performed based on structural masks, and the various structures of the heart are presented using a semi-transparent rendering method. The spatial distribution characteristics of the indentation area are highlighted or shown in solid form. Simultaneously, interactive functions such as multi-angle observation and distance measurement are provided to assist clinicians in assessing the morphology and potential risks of the indentation area from multiple perspectives.

[0097] To achieve the overlay display of various data in a unified coordinate system, a rigid transformation matrix is ​​used to map the coordinates of the CT image, structural mask, and indentation curve. The transformation formula is as follows:

[0098] Where source∈{CT,mask,curve}

[0099] Among them, T source It is a rigid transformation matrix, defined as:

[0100]

[0101] R is the rotation matrix, t is the translation vector, (x source ,y source ,z source The coordinates of the original data are used. Through the above transformation, the CT image, mask, and spatial curve are accurately superimposed and visualized in the same coordinate system.

Claims

1. An AI-driven method for three-dimensional analysis of left atrial CT and automatic assessment of indentation areas, characterized in that, include: Step 1: Acquire three-dimensional medical image data of the patient's heart; Step 2: Based on deep learning models or image segmentation algorithms, extract the relevant structures of the left atrium from the 3D medical image data, including the left atrium, left atrial wall, esophagus, coronary sinus and aorta; Step 3: Based on the anatomical structure segmentation results obtained in Step 2, extract the three-dimensional label map of each anatomical structure and construct the label mask of the left atrial wall and adjacent organs; using the label mask of the left atrial wall as the reference area, calculate the voxel-level spatial distance with the target areas represented by the label masks of the aorta, coronary sinus and esophagus respectively. A distance threshold is set, and left atrial wall voxel points whose shortest spatial distance to the target area is less than the distance threshold are selected as the indentation area point set, thereby automatically locating the indentation area and extracting the point set. Step 4: Extract and quantify the geometric contour of the identified left atrial indentation area; Step 5: Visualization of the results of multiple indentations in the left atrium.

2. The method for three-dimensional analysis of left atrial CT and automatic assessment of indentation area according to claim 1, wherein in step 1, the three-dimensional medical image data is medical image data containing complete left ventricular structure information of the human body, including CT images or MRI images.

3. The method for three-dimensional analysis of left atrial CT and automatic assessment of indentation area according to claim 2, in step 2, medical software is used to manually annotate the structure, or artificial intelligence deep learning algorithm is used to automatically complete the segmentation of the main anatomical structures of the left atrial region. The deep learning algorithm includes STU-Net-B and SwinUNETR. The segmented left atrial region consists of the left atrium, left atrial wall, coronary sinus, aorta and esophagus in sequence.

4. The method for three-dimensional analysis of left atrial CT and automatic assessment of indentation area according to claim 3, wherein step 3 further includes the following steps: Step 3.1: The label map is a three-dimensional voxel grid data structure. Each voxel position has a corresponding label value and a three-dimensional spatial coordinate index. The left atrial wall, aorta, coronary sinus and esophagus are assigned different label values. In order to improve the accuracy and stability of subsequent calculations, the segmentation results are preprocessed, including removing isolated noise points and filling small holes, so as to ensure the boundary continuity and structural integrity of the label area. Step 3.2: For each pair of voxels between the reference region and the target region, calculate the shortest spatial distance using the fast Euclidean distance transformation algorithm. The specific calculation method is as follows: in, and These are the spatial coordinates of any two voxels in any two label images; the threshold is 2 mm; to improve calculation and filtering efficiency, all label voxels are spatially indexed based on the KD tree data structure to accelerate the search process for nearest neighbor voxels; Step 3.3: Perform contour extraction and geometric optimization on the acquired indentation area point set; first, project the 3D point set onto a preset 2D plane to obtain the projected 2D point set of the indentation area: in, Indicates the coordinates of the indentation point on a two-dimensional plane; Based on this point set, the Alpha-shape algorithm is applied to extract the boundary contour of the indentation region. The contour is defined as follows: in It is a parameter that controls the complexity of the boundary shape, and its value affects the degree of contour shrinkage and the effect of local feature preservation. After obtaining the initial contour, the boundary contour is further geometrically optimized: First, the contour is simplified using the Douglas-Peucker algorithm, removing redundant vertices while maintaining the main geometric features of the contour; then, B-spline curves are applied to smooth the simplified contour to improve its continuity and smoothness, providing optimized boundary input for subsequent quantitative analysis.

5. The method for three-dimensional analysis of left atrial CT and automatic assessment of indentation area according to claim 4, wherein step 4 includes the following steps: Step 4.1: Based on the extracted indentation contour data, calculate the perimeter of the indentation area; the contour is usually composed of several straight line segments and curved lines; in order to calculate the perimeter, the length of each line segment needs to be summed first; for straight line segments, the length is directly calculated based on the coordinates of their endpoints. For curved segments, a piecewise linear approximation method is used to approximate them as multiple short line segments to improve the accuracy of the calculation. The total perimeter of the indentation area is obtained by summing the lengths of all contour line segments. The key to this process is to perform accurate approximation calculations on the curved parts to ensure the accuracy of the overall perimeter. Step 4.2: Since the outline of the indentation area is a closed curve in three-dimensional space, the area calculation method of three-dimensional polygonal patches is used for processing. Specifically, the spatial surface enclosed by the outline is triangulated into multiple triangular patches, the area of ​​each triangular patch is calculated, and the areas are summed to obtain the total area of ​​the entire indentation area.

6. The method for three-dimensional analysis of left atrial CT and automatic assessment of indentation area according to claim 5, in step 5, firstly, the original CT image and the automatically segmented structural mask are registered and superimposed to accurately display the boundaries of each organ; then, the mask is reconstructed in three dimensions using three-dimensional reconstruction technology, so that the indentation area is clearly presented in a structured form in three-dimensional space; finally, the indentation area, CT image and reconstructed mask are comprehensively superimposed to intuitively present the spatial location, morphology and characteristics of the indentation and its relationship with the surrounding anatomical structures, and comprehensively assist in the assessment of the distribution and risk characteristics of the indentation area from multiple perspectives.

7. An AI-driven left atrial CT three-dimensional analysis and automatic assessment system for indentation region for implementing the method of any one of claims 1 to 6, the system comprising: The medical image acquisition module is used to acquire three-dimensional medical image data containing the left atrial region; The structural segmentation module is used to identify anatomical structures in the image data and extract the left atrium, left atrial wall, esophagus, aorta and coronary sinus, wherein the structural segmentation is achieved by an artificial intelligence deep learning model or a medical image processing algorithm. The indentation localization module is used to spatially locate and extract point sets of the indentation area between the left atrial wall and adjacent organs based on the label map and spatial distance calculation strategy. The geometric calculation module is used to extract the contour, calculate the perimeter and calculate the area of ​​the indentation area. The contour extraction includes Alpha-shape contour construction, Douglas-Peucker algorithm simplification and B-spline curve smoothing. The area calculation includes dividing the projected surface of the indentation area into multiple triangular patches and calculating the area of ​​each patch. The visualization module is used to fuse the indentation area with CT images and structural masks to present the spatial location, geometric shape, and relative relationship of the indentation to the anatomical structure, which is used to assist in preoperative clinical assessment.

Images