Lung tumor cell diffusion direction marking method and device and computer equipment

Abstract

The invention relates to a lung tumor cell diffusion direction marking method and device and computer equipment. The method comprises the following steps: constructing a lymph and blood vessel framework diagram, and identifying a blood vessel structure and a lymph structure in a lung CT image to be marked by adopting a pre-trained machine learning model; registration is carried out according to the lymph and blood vessel framework diagram, and a registered blood vessel structure and a registered lymph structure are obtained; identifying a tumor in the registered lung CT image, a first position relation between the tumor and a registered vascular structure, a second position relation between the tumor and a registered lymph structure, and a third position relation between the tumor in a lung structure; according to the first position relation, the second position relation and the third position relation, the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor are calculated and marked. By adopting the method, accurate and efficient diffusion marking can be realized.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device and computer equipment for marking the diffusion direction of lung tumor cells. Background Art

[0002] The spread of lung tumors is one of the most critical challenges in cancer treatment and management, and is also one of the means to assess whether there is a risk of invasion to other organs. Traditional tumor spread assessment relies on biomarkers and imaging methods such as computed tomography (CT) and magnetic resonance imaging (MRI). Although these methods provide detailed information about the location, size, and morphology of the tumor, they usually cannot directly reveal the spread path and direction of tumor cells.

[0003] At present, the marking of the direction of tumor spread mainly relies on detailed image analysis combined with clinical pathological data. Using advanced image processing software, doctors can observe and analyze the relationship between the tumor and surrounding structures in three-dimensional space, and try to determine the possible diffusion path. In addition, through dynamic contrast-enhanced imaging technology (such as dynamic contrast-enhanced MRI), attempts are made to obtain data on the dynamic changes of tumor blood flow to infer the activity level and potential diffusion tendency of the tumor. However, these technologies require a high degree of expertise and experience, and are complex to operate and highly dependent. In addition, imaging technology can often only capture structural changes that have already occurred, rather than real-time dynamic diffusion processes, which limits its application in early prediction and intervention. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device and computer equipment for marking the diffusion direction of lung tumor cells in response to the above technical problems.

[0005] A method for marking the diffusion direction of lung tumor cells, the method comprising:

[0006] Extracting a vascular system image and a lymphatic system image from CT image data of the abnormal lung, converting the vascular system image and the lymphatic system image into 3D using a skeletonization technique to obtain a vascular system framework diagram and a lymphatic system framework diagram, and performing 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain a lymphatic and vascular framework diagram;

[0007] Using a pre-trained machine learning model to identify a lung CT image to be labeled, and obtaining a vascular structure and a lymphatic structure in the lung CT image;

[0008] According to the lymphatic and vascular framework diagram, aligning the vascular structure and the lymphatic structure in the lung CT image to obtain an aligned vascular structure and an aligned lymphatic structure;

[0009] Identifying a tumor in the registered lung CT image, a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure;

[0010] Based on the first positional relationship, the second positional relationship and the third positional relationship, the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor are calculated, and the diffusion direction of the lung tumor cells is marked based on the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion.

[0011] In one of the embodiments, it also includes: acquiring original CT scan data using a contrast agent, reconstructing the original CT scan data through multi-planar reconstruction to obtain the CT image data; the visibility of lymphatic structures and blood vessels in the CT image data is greater than a threshold.

[0012] In one of the embodiments, it also includes: extracting vascular lines and lymphatic lines of the vascular structure and lymphatic structure in the vascular system image and the lymphatic system image to obtain the vascular structure topology and the lymphatic structure topology; smoothing the vascular structure topology and the lymphatic structure topology to obtain a smooth vascular structure topology and a smooth lymphatic structure topology; marking the structural nodes in the smooth vascular structure topology and the smooth lymphatic structure topology, and adjusting the smooth vascular structure topology and the smooth lymphatic structure topology according to the preset skeleton thickness parameter, curvature parameter and the structural nodes to obtain a 3D vascular system framework diagram and a lymphatic system framework diagram.

[0013] In one of the embodiments, the method further includes: using a pre-trained U-Net model to identify lung CT images to be labeled.

[0014] In one of the embodiments, it also includes: using the SIFT algorithm to extract key points in the vascular structure and lymphatic structure in the lung CT image, and obtaining descriptors of the key points; the descriptors at least include: thickness parameters and curvature parameters; according to the descriptors, matching the structure nodes with the key points; according to the matched key points, aligning the vascular structure and lymphatic structure in the lung CT image to obtain aligned vascular structures and aligned lymphatic structures.

[0015] In one of the embodiments, it also includes: identifying a tumor in the registered lung CT image, extracting edge information of the tumor, obtaining contour information of the tumor, and constructing a first relationship function of a first position relationship, a second relationship function of a second position relationship, and a third relationship function of a third position relationship in the coordinate system according to the contour information and the registered vascular structure; the first relationship function determines the vascular diffusion probability and the vascular diffusion direction according to the distance and pointing direction from the midpoint of the contour information to the point on the registered vascular structure; the second relationship function determines the lymphatic diffusion probability and the lymphatic diffusion direction according to the distance and pointing direction from the midpoint of the contour information to the point on the registered lymphatic structure; the third relationship function determines the direct diffusion probability and the direct diffusion direction according to the distance and pointing direction from the midpoint of the contour information to the point on the edge of the lung structure.

[0016] In one embodiment, the first relationship function includes:

[0017] f v (p)=(q∈Vmin||pq||,atan2(q y -p y ,q x -p x ))

[0018] Where V represents the point set in the registered vascular structure, atan2 is the angle calculation function, p represents the midpoint of the contour information, and q represents the point in the registered vascular structure; f v (p) represents the angle relationship in the first relationship function;

[0019] The second relationship function is:

[0020] f l (p)=(q∈Lmin||pq||,atan2(q y -p y ,q x -p x ))

[0021] L represents the point set in the lung structure;

[0022] The third relationship function is:

[0023] f d (p)=(q∈Dmin||pq||,atan2(q y -p y ,q x -p x ))

[0024] D represents the point in the registered lymphatic structure;

[0025] The diffusion probability models of the first relationship function, the second relationship function and the third relationship function are:

[0026] P(d)=e -αd

[0027] d represents the distance of the nearest point, and α is a parameter adjusted based on clinical data; and further comprising: using the first relationship function, the second relationship function, the third relationship function and the diffusion probability model, traversing the midpoint of the contour information to obtain the first relationship data f v (P, θ), the second relational data f l (P,θ) and the third relation data f d (P,θ), P represents the probability of the midpoint of the contour information being at the diffusion angle θ;

[0028] For any diffusion angle, the first relationship data, the second relationship data and the third relationship data are fused to obtain the diffusion direction and probability of lung tumor cells.

[0029] A lung tumor cell diffusion direction marking device, the device comprising:

[0030] A framework diagram construction module is used to extract a vascular system image and a lymphatic system image from CT image data of abnormal lungs, convert the vascular system image and the lymphatic system image into 3D images using skeletonization technology to obtain a vascular system framework diagram and a lymphatic system framework diagram, and perform 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain lymphatic and vascular framework diagrams;

[0031] A recognition module, used to use a pre-trained machine learning model to recognize a lung CT image to be labeled, and obtain a vascular structure and a lymphatic structure in the lung CT image;

[0032] A registration module, used for registering the vascular structure and the lymphatic structure in the lung CT image according to the lymphatic and vascular framework diagram to obtain a registered vascular structure and a registered lymphatic structure;

[0033] a positional relationship confirmation module, configured to identify a tumor in the registered lung CT image, a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure;

[0034] The marking module is used to calculate the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor according to the first position relationship, the second position relationship and the third position relationship, and mark the diffusion direction of lung tumor cells according to the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion.

[0035] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] Extracting a vascular system image and a lymphatic system image from CT image data of the abnormal lung, converting the vascular system image and the lymphatic system image into 3D using a skeletonization technique to obtain a vascular system framework diagram and a lymphatic system framework diagram, and performing 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain a lymphatic and vascular framework diagram;

[0037] Using a pre-trained machine learning model to identify a lung CT image to be labeled, and obtaining a vascular structure and a lymphatic structure in the lung CT image;

[0038] According to the lymphatic and vascular framework diagram, aligning the vascular structure and the lymphatic structure in the lung CT image to obtain an aligned vascular structure and an aligned lymphatic structure;

[0039] Identifying a tumor in the registered lung CT image, a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure;

[0040] Based on the first positional relationship, the second positional relationship and the third positional relationship, the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor are calculated, and the diffusion direction of the lung tumor cells is marked based on the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0042] Extracting a vascular system image and a lymphatic system image from CT image data of the abnormal lung, converting the vascular system image and the lymphatic system image into 3D using a skeletonization technique to obtain a vascular system framework diagram and a lymphatic system framework diagram, and performing 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain a lymphatic and vascular framework diagram;

[0043] Using a pre-trained machine learning model to identify a lung CT image to be labeled, and obtaining a vascular structure and a lymphatic structure in the lung CT image;

[0044] According to the lymphatic and vascular framework diagram, aligning the vascular structure and the lymphatic structure in the lung CT image to obtain an aligned vascular structure and an aligned lymphatic structure;

[0045] Identifying a tumor in the registered lung CT image, a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure;

[0046] Based on the first positional relationship, the second positional relationship and the third positional relationship, the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor are calculated, and the diffusion direction of the lung tumor cells is marked based on the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion.

[0047] The above-mentioned lung tumor cell diffusion direction marking method, device, computer equipment and storage medium significantly improve the recognition and prediction accuracy of lung tumor diffusion paths by integrating advanced medical image processing, three-dimensional modeling and machine learning technologies. First, skeletonization technology is used to extract and construct a three-dimensional framework diagram of the vascular and lymphatic systems from abnormal lung CT images. Then, the new CT image is analyzed by a pre-trained machine learning model to automatically identify the vascular and lymphatic structures and accurately align them with the pre-built framework diagram. Such alignment makes it possible to accurately locate the position of the tumor relative to key physiological structures, and further evaluate the probability and direction of tumor cell diffusion along blood vessels, lymph or direct invasion by calculating the spatial relationship with each structure. The present invention provides a highly accurate method to predict and mark tumor diffusion paths, which provides important support for clinical decision-making and helps to formulate more effective treatment strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of a process for marking the diffusion direction of lung tumor cells in one embodiment;

[0049] Figure 2 is a structural block diagram of a lung tumor cell diffusion direction marking device in one embodiment;

[0050] Figure 3 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0052] In one embodiment, Figure 1 As shown, a method for marking the diffusion direction of lung tumor cells is provided, comprising the following steps:

[0053] Step 102, extracting the vascular system image and the lymphatic system image from the CT image data of the abnormal lung, converting the vascular system image and the lymphatic system image into 3D using skeletonization technology to obtain a vascular system framework diagram and a lymphatic system framework diagram, and performing 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain lymphatic and vascular framework diagrams.

[0054] In this step, a large number of samples are used to construct lymphatic and vascular framework diagrams, which can provide important support for the subsequent decision-making. In addition, a standard coordinate system is constructed when modeling the framework diagram, which provides a basis for subsequent calculations.

[0055] Step 104: Use a pre-trained machine learning model to identify the lung CT image to be labeled to obtain the vascular structure and lymphatic structure in the lung CT image.

[0056] For the lung CT images to be labeled, a pre-trained machine learning model can be used to identify the vascular structure and lymphatic structure.

[0057] Step 106 , registering the vascular structure and the lymphatic structure in the lung CT image according to the lymphatic and vascular framework diagram to obtain a registered vascular structure and a registered lymphatic structure.

[0058] It can be seen that since the conditions of lung CT images are difficult to standardize, effective results cannot be obtained by relying solely on machine learning model recognition. In this step, accurate registration of vascular structures and lymphatic structures can be obtained through registration.

[0059] Step 108 , identifying a tumor in the registered lung CT image, as well as a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure.

[0060] Step 110, based on the first positional relationship, the second positional relationship and the third positional relationship, the probability and diffusion direction of the tumor cells in the tumor to diffuse into the blood vessels, lymphatic diffusion and direct diffusion are calculated, and the diffusion direction of the lung tumor cells is marked according to the probability and diffusion direction of the blood vessels, lymphatic diffusion and direct diffusion.

[0061] In the above-mentioned method for marking the direction of lung tumor cell diffusion, the accuracy of identifying and predicting the path of lung tumor diffusion is significantly improved by integrating advanced medical image processing, three-dimensional modeling and machine learning technologies. First, skeletonization technology is used to extract and construct a three-dimensional framework diagram of the vascular and lymphatic systems from abnormal lung CT images. Then, the new CT image is analyzed by a pre-trained machine learning model to automatically identify the vascular and lymphatic structures and accurately align them with the pre-built framework diagram. Such alignment makes it possible to accurately locate the position of the tumor relative to key physiological structures, and further evaluate the probability and direction of tumor cell diffusion along blood vessels, lymph or direct invasion by calculating the spatial relationship with each structure. The present invention provides a highly accurate method to predict and mark the path of tumor diffusion, which provides important support for clinical decision-making and helps to formulate more effective treatment strategies.

[0062] In one of the embodiments, raw CT scan data using a contrast agent is acquired, and the raw CT scan data is reconstructed through multi-planar reconstruction to obtain CT image data; the visibility of lymphatic structures and blood vessels in the CT image data is greater than a threshold.

[0063] In another embodiment, vascular lines and lymphatic lines of vascular structures and lymphatic structures in vascular system images and lymphatic system images are extracted to obtain vascular structure topology and lymphatic structure topology; the vascular structure topology and lymphatic structure topology are smoothed to obtain smooth vascular structure topology and smooth lymphatic structure topology; structural nodes in the smooth vascular structure topology and smooth lymphatic structure topology are marked, and the smooth vascular structure topology and smooth lymphatic structure topology are adjusted according to pre-set skeleton thickness parameters, curvature parameters and structural nodes to obtain a 3D vascular system framework diagram and a lymphatic system framework diagram.

[0064] In one of the embodiments, a pre-trained U-Net model is used to identify lung CT images to be labeled.

[0065] In another embodiment, a SIFT algorithm is used to extract key points in the vascular structure and lymphatic structure in the lung CT image, and to obtain descriptors of the key points; the descriptors include at least: a thickness parameter and a curvature parameter; according to the descriptors, the structure nodes are matched with the key points; according to the matched key points, the vascular structure and lymphatic structure in the lung CT image are registered to obtain registered vascular structures and registered lymphatic structures.

[0066] In one of the embodiments, a tumor is identified in the registered lung CT image, edge information of the tumor is extracted, contour information of the tumor is obtained, and a first relationship function of a first position relationship, a second relationship function of a second position relationship, and a third relationship function of a third position relationship are constructed in the coordinate system based on the contour information and the registered vascular structure; the first relationship function determines the probability of vascular diffusion and the direction of vascular diffusion based on the distance and pointing direction from the midpoint of the contour information to a point on the registered vascular structure; the second relationship function determines the probability of lymphatic diffusion and the direction of lymphatic diffusion based on the distance and pointing direction from the midpoint of the contour information to a point on the registered lymphatic structure; the third relationship function determines the probability of direct diffusion and the direction of direct diffusion based on the distance and pointing direction from the midpoint of the contour information to a point on the edge of the lung structure.

[0067] In one embodiment, the first relationship function includes:

[0068] f v (p)=(q∈Vmin||pq||,atan2(q y -p y ,q x -p x ))

[0069] Where V represents the point set in the registered vascular structure, atan2 is the angle calculation function, p represents the midpoint of the contour information, and q represents the point in the registered vascular structure; f v (p) represents the angle relationship in the first relationship function;

[0070] The second relationship function is:

[0071] f l (p)=(q∈Lmin||pq||,atan2(q y -p y ,q x -p x ))

[0072] L represents the point set in the lung structure;

[0073] The third relationship function is:

[0074] f d (p)=(q∈Dmin||pq||,atan2(q y -p y ,q x -p x ))

[0075] D represents the point in the registered lymphatic structure;

[0076] The diffusion probability models of the first relationship function, the second relationship function and the third relationship function are:

[0077] P(d)=e -αd

[0078] d represents the distance to the closest point, and α is a parameter adjusted based on clinical data;

[0079] Calculating the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor according to the first position relationship, the second position relationship and the third position relationship includes: using the first relationship function, the second relationship function, the third relationship function and the diffusion probability model, traversing the midpoint of the contour information, and obtaining the first relationship data f v (P, θ), the second relational data f l (P,θ) and the third relation data f d (P, θ), where P represents the probability that the midpoint of the contour information is at the diffusion angle θ; for any diffusion angle, the first relationship data, the second relationship data, and the third relationship data are fused to obtain the diffusion direction and probability of lung tumor cells.

[0080] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0081] In one embodiment, Figure 2 As shown, a lung tumor cell diffusion direction marking device is provided, comprising a framework map construction module 202, a recognition module 204, a registration module 206, a position relationship confirmation module 208 and a marking module 210, wherein:

[0082] A framework diagram construction module 202 is used to extract a vascular system image and a lymphatic system image from the CT image data of the abnormal lung, convert the vascular system image and the lymphatic system image into 3D using a skeletonization technique to obtain a vascular system framework diagram and a lymphatic system framework diagram, and perform 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using a 3D modeling software to obtain a lymphatic and vascular framework diagram;

[0083] An identification module 204 is used to identify the lung CT image to be labeled using a pre-trained machine learning model to obtain the vascular structure and lymphatic structure in the lung CT image;

[0084] A registration module 206, configured to register the vascular structure and the lymphatic structure in the lung CT image according to the lymphatic and vascular framework map, to obtain a registered vascular structure and a registered lymphatic structure;

[0085] A positional relationship confirmation module 208, for identifying a tumor in the registered lung CT image, and a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure;

[0086] The marking module 210 is used to calculate the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor based on the first position relationship, the second position relationship and the third position relationship, and mark the diffusion direction of the lung tumor cells based on the probability and diffusion direction of the vascular diffusion, lymphatic diffusion and direct diffusion.

[0087] In one of the embodiments, the framework diagram construction module 202 is also used to obtain raw CT scan data using a contrast agent, and reconstruct the raw CT scan data through multi-planar reconstruction to obtain the CT image data; the visibility of lymphatic structures and blood vessels in the CT image data is greater than a threshold.

[0088] In one of the embodiments, the framework diagram construction module 202 is also used to extract the vascular lines and lymphatic lines of the vascular structure and lymphatic structure in the vascular system image and the lymphatic system image, so as to obtain the vascular structure topology and the lymphatic structure topology; smooth the vascular structure topology and the lymphatic structure topology to obtain a smooth vascular structure topology and a smooth lymphatic structure topology; mark the structural nodes in the smooth vascular structure topology and the smooth lymphatic structure topology, and adjust the smooth vascular structure topology and the smooth lymphatic structure topology according to the preset skeleton thickness parameter, curvature parameter and the structural nodes to obtain a 3D vascular system framework diagram and a lymphatic system framework diagram.

[0089] In one embodiment, the recognition module 204 is further configured to use a pre-trained U-Net model to recognize the lung CT image to be labeled.

[0090] In one of the embodiments, the registration module 206 is further used to extract key points in the vascular structure and lymphatic structure in the lung CT image using the SIFT algorithm, and obtain descriptors of the key points; the descriptors include at least: thickness parameters and curvature parameters; according to the descriptors, the structure nodes are matched with the key points; according to the matched key points, the vascular structure and lymphatic structure in the lung CT image are registered to obtain registered vascular structures and registered lymphatic structures.

[0091] In one embodiment, the position relationship confirmation module 208 is also used to identify the tumor in the registered lung CT image, extract the edge information of the tumor, obtain the contour information of the tumor, and construct a first relationship function of the first position relationship, a second relationship function of the second position relationship, and a third relationship function of the third position relationship in the coordinate system according to the contour information and the registered vascular structure; the first relationship function determines the vascular diffusion probability and the vascular diffusion direction according to the distance and pointing direction from the midpoint of the contour information to the point on the registered vascular structure; the second relationship function determines the lymphatic diffusion probability and the lymphatic diffusion direction according to the distance and pointing direction from the midpoint of the contour information to the point on the registered lymphatic structure; the third relationship function determines the direct diffusion probability and the direct diffusion direction according to the distance and pointing direction from the midpoint of the contour information to the point on the edge of the lung structure.

[0092] In one embodiment, the first relationship function includes:

[0093] f v (p)=(q∈Vmin||pq||,atan2(q y -p y ,q x -p x ))

[0094] Where V represents the point set in the registered vascular structure, atan2 is the angle calculation function, p represents the midpoint of the contour information, and q represents the point in the registered vascular structure; f v (p) represents the angle relationship in the first relationship function;

[0095] The second relationship function is:

[0096] f l (p)=(q∈Lmin||pq||,atan2(q y -p y ,q x -p x ))

[0097] L represents the point set in the lung structure;

[0098] The third relationship function is:

[0099] f d (p)=(q∈Dmin||pq||,atan2(q y -p y ,q x -p x ))

[0100] D represents the point in the registered lymphatic structure;

[0101] The diffusion probability models of the first relationship function, the second relationship function and the third relationship function are:

[0102] P(d)=e -αd

[0103] d represents the distance of the nearest point, and α is a parameter adjusted based on clinical data; using the first relationship function, the second relationship function, the third relationship function and the diffusion probability model, the marking module 210 is also used to traverse the midpoint of the contour information to obtain the first relationship data f v (P, θ), the second relational data f l (P,θ) and the third relation data f d (P, θ), where P represents the probability that the midpoint of the contour information is at the diffusion angle θ; for any diffusion angle, the first relationship data, the second relationship data, and the third relationship data are fused to obtain the diffusion direction and probability of lung tumor cells.

[0104] For the specific definition of the lung tumor cell diffusion direction marking device, please refer to the definition of the lung tumor cell diffusion direction marking method mentioned above, which will not be repeated here. Each module in the above lung tumor cell diffusion direction marking device can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0105] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 3As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for marking the diffusion direction of lung tumor cells is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0106] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0107] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in the above embodiment when executing the computer program.

[0108] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in the above embodiment are implemented.

[0109] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0110] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for marking the diffusion direction of lung tumor cells, characterized in that: The method comprises: Extracting a vascular system image and a lymphatic system image from CT image data of the abnormal lung, converting the vascular system image and the lymphatic system image into 3D using a skeletonization technique to obtain a vascular system framework diagram and a lymphatic system framework diagram, and performing 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain a lymphatic and vascular framework diagram; Using a pre-trained machine learning model to identify a lung CT image to be labeled, and obtaining a vascular structure and a lymphatic structure in the lung CT image; According to the lymphatic and vascular framework diagram, aligning the vascular structure and the lymphatic structure in the lung CT image to obtain an aligned vascular structure and an aligned lymphatic structure; Identifying a tumor in the registered lung CT image, a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure; Based on the first positional relationship, the second positional relationship and the third positional relationship, the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor are calculated, and the diffusion direction of the lung tumor cells is marked based on the probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion.

2. The method according to claim 1, characterized in that The steps of acquiring CT image data include: Original CT scan data using a contrast agent is acquired, and the original CT scan data is reconstructed through multi-planar reconstruction to obtain the CT image data; the visibility of lymphatic structures and blood vessels in the CT image data is greater than a threshold value.

3. The method according to claim 1, characterized in that: The vascular system image and the lymphatic system image are converted into 3D images by using skeletonization technology to obtain a vascular system framework diagram and a lymphatic system framework diagram, including: Extracting vascular lines and lymphatic lines of the vascular structure and lymphatic structure in the vascular system image and the lymphatic system image, thereby obtaining a vascular structure topology and a lymphatic structure topology; Smoothing the vascular structure topology and the lymphatic structure topology to obtain a smoothed vascular structure topology and a smoothed lymphatic structure topology; The structural nodes in the smooth vascular structure topology and the smooth lymphatic structure topology are marked, and the smooth vascular structure topology and the smooth lymphatic structure topology are adjusted according to the preset skeleton thickness parameter, curvature parameter and the structural nodes to obtain a 3D vascular system framework diagram and a lymphatic system framework diagram.

4. The method according to claim 1, characterized in that: Use pre-trained machine learning models to identify lung CT images to be labeled, including: The pre-trained U-Net model is used to identify the lung CT images to be labeled.

5. The method according to claim 3, characterized in that: According to the lymphatic and vascular framework diagram, the vascular structure and the lymphatic structure in the lung CT image are registered to obtain the registered vascular structure and the registered lymphatic structure, including: The SIFT algorithm is used to extract key points in the vascular structure and lymphatic structure in the lung CT image, and to obtain descriptors of the key points; the descriptors at least include: a thickness parameter and a curvature parameter; According to the descriptor, matching the structural node with the key point; According to the key points obtained by matching, the vascular structure and the lymphatic structure in the lung CT image are registered to obtain a registered vascular structure and a registered lymphatic structure.

6. The method according to any one of claims 1 to 5, characterized in that: Identifying a tumor in the registered lung CT image, and acquiring a first positional relationship between the tumor and the registered blood vessel structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure, including: A tumor is identified in the registered lung CT image, edge information of the tumor is extracted, contour information of the tumor is obtained, and a first relationship function of a first position relationship, a second relationship function of a second position relationship, and a third relationship function of a third position relationship are constructed in the coordinate system according to the contour information and the registered vascular structure; the first relationship function determines the probability of vascular diffusion and the direction of vascular diffusion according to the distance and pointing direction from the midpoint of the contour information to the point on the registered vascular structure; the second relationship function determines the probability of lymphatic diffusion and the direction of lymphatic diffusion according to the distance and pointing direction from the midpoint of the contour information to the point on the registered lymphatic structure; the third relationship function determines the probability of direct diffusion and the direction of direct diffusion according to the distance and pointing direction from the midpoint of the contour information to the point on the edge of the lung structure.

7. The method according to claim 6, characterized in that The first relationship function includes: f v (p)=(q∈Vmin||p-q||,atan2(q y -p y ,q x -p x )) Where V represents the point set in the registered vascular structure, atan2 is the angle calculation function, p represents the midpoint of the contour information, and q represents the point in the registered vascular structure; f v (p) represents the angle relationship in the first relationship function; The second relationship function is: f l (p)=(q∈Lmin||p-q||,atan2(q y -p y ,q x -p x )) L represents the point set in the lung structure; The third relationship function is: f d (p)=(q∈Dmin||p-q||,atan2(q y -p y ,q x -p x )) D represents the points in the registered lymphatic structure; The diffusion probability models of the first relationship function, the second relationship function and the third relationship function are: P(d)=e -αd d represents the distance to the closest point, and α is a parameter adjusted based on clinical data; The calculated probabilities and diffusion directions of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor according to the first position relationship, the second position relationship and the third position relationship include: The first relationship function, the second relationship function, the third relationship function and the diffusion probability model are used to traverse the midpoints of the contour information to obtain the first relationship data f v (P, θ), the second relational data f l (P,θ) and the third relation data f d (P,θ), P represents the probability of the midpoint of the contour information being at the diffusion angle θ; For any diffusion angle, the first relationship data, the second relationship data and the third relationship data are fused to obtain the diffusion direction and probability of lung tumor cells.

8. A device for marking the diffusion direction of lung tumor cells, characterized in that: The device comprises: A framework diagram construction module is used to extract a vascular system image and a lymphatic system image from CT image data of abnormal lungs, convert the vascular system image and the lymphatic system image into 3D images using skeletonization technology to obtain a vascular system framework diagram and a lymphatic system framework diagram, and perform 3D modeling on the vascular system framework diagram and the lymphatic system framework diagram in a coordinate system using 3D modeling software to obtain lymphatic and vascular framework diagrams; A recognition module, used to use a pre-trained machine learning model to recognize a lung CT image to be labeled, and obtain a vascular structure and a lymphatic structure in the lung CT image; A registration module, used for registering the vascular structure and the lymphatic structure in the lung CT image according to the lymphatic and vascular framework diagram to obtain a registered vascular structure and a registered lymphatic structure; a positional relationship confirmation module, configured to identify a tumor in the registered lung CT image, a first positional relationship between the tumor and the registered vascular structure, a second positional relationship between the tumor and the registered lymphatic structure, and a third positional relationship between the tumor and the lung structure; The marking module is used to calculate the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion of tumor cells in the tumor according to the first position relationship, the second position relationship and the third position relationship, and mark the diffusion direction of lung tumor cells according to the probability and diffusion direction of vascular diffusion, lymphatic diffusion and direct diffusion.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images