Pancreatic lesion resection procedure assistance system, method, and apparatus

Abstract

The application relates to a pancreatic lesion resection type assisted decision system, method, device, computer equipment and storage medium. The system comprises a processor which performs: acquiring a pancreatic three-dimensional medical image, the pancreatic three-dimensional medical image comprising a lesion; identifying the number of lesions and the nature of the lesions, the number of lesions being one or more, and the nature of the lesions comprising at least one of a first lesion type and a second lesion type; and constructing a decision tree model according to the number of lesions, the nature of the lesions and a surgical function to generate an assisted decision of a resection type corresponding to the lesion. The system can fully consider the specificity of different pancreatic lesions, construct a decision tree model of the resection type of the lesion according to the number of lesions, the nature of the lesions and the surgical function, generate an assisted decision of the resection type corresponding to the lesion according to the specific conditions of different lesions, and use the judgment branches of the decision tree model to accurately, efficiently and intelligently assist doctors in selecting the resection type of the pancreatic lesion.

Description

Technical Field

[0001] This application relates to the field of medical imaging technology, and in particular to a decision support system, method, device, computer equipment, and storage medium for pancreatic lesion resection surgery. Background Technology

[0002] Pancreatic tumors are highly malignant, and surgical resection is an effective treatment. CT (computed tomography) and MRI (magnetic resonance imaging) images are crucial for assessing the condition and planning preoperative surgery for pancreatic tumors. Currently, surgeons typically assess the condition by observing the location and relationships of organs and tissues such as the pancreas, pancreatic duct, tumor, and blood vessels layer by layer on two-dimensional images, and then evaluating the tumor resection method through subjective assessment or two-dimensional ranging tools. However, pancreatic tumors often progress rapidly, making this method less intuitive, requiring a high level of experience from the surgeon, and exhibiting low accuracy and slow efficiency. Summary of the Invention

[0003] Therefore, it is necessary to provide a pancreatic lesion resection surgical auxiliary decision-making system, method, device, computer equipment, and storage medium to address the above-mentioned technical problems.

[0004] One embodiment of this specification provides a pancreatic lesion resection surgical procedure auxiliary decision-making system. The system includes a processor, which is configured to perform the following steps: acquiring three-dimensional medical images of the pancreas, the pancreatic three-dimensional medical images including lesions; identifying the number and nature of lesions, the number of lesions being one or more, the nature of the lesions including at least one of a first lesion type and a second lesion type; constructing a decision tree model based on the number of lesions, the nature of the lesions, and a surgical procedure function to generate an auxiliary decision for the resection surgical procedure corresponding to the lesion.

[0005] In one embodiment, when the processor executes the acquisition of three-dimensional pancreatic medical images, it further includes performing the following steps: acquiring medical images of multiple modalities, including a first modality and a second modality; acquiring medical images at multiple time phases in the first modality, and registering the multiple time phases of the medical images acquired in the first modality to obtain a first-modality pancreatic medical image; acquiring medical images at multiple time sequences in the second modality, and registering the multiple time sequences of the medical images acquired in the second modality to obtain a second-modality pancreatic medical image; registering the first-modality pancreatic medical image and the second-modality pancreatic medical image to obtain a registered pancreatic medical image; and performing three-dimensional reconstruction on the registered pancreatic medical image to obtain the three-dimensional pancreatic medical image.

[0006] In one embodiment, when the processor executes the step of constructing a decision tree model based on the number of lesions, the nature of the lesions, and the surgical procedure function to generate an auxiliary decision on the resection procedure corresponding to the lesion, the step further includes the following steps: constructing the surgical procedure function based on lesion parameters, wherein the lesion parameters include at least one of lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel; the surgical procedure function includes a first surgical procedure function, a second surgical procedure function, and a third surgical procedure function; constructing the root node of the decision tree model based on the number of lesions; when the number of lesions is one, constructing the first and second level nodes of the decision tree model based on the nature of the lesions; and constructing the decision tree model based on the first surgical procedure function. The decision tree model is constructed using the first and third level nodes of the model, or the second and third level nodes of the decision tree model constructed according to the second surgical technique function. The first surgical technique matching engine is used to generate an auxiliary decision on the resection procedure corresponding to the lesion. The first surgical technique matching engine is used to characterize the mapping relationship between the output value of the first surgical technique function or the output value of the second surgical technique function and the resection procedure. The root node of the decision tree model is constructed according to the number of lesions. When there are multiple lesions, the second and second level nodes of the decision tree model are constructed according to the third surgical technique function. The second surgical technique matching engine is used to generate an auxiliary decision on the resection procedure corresponding to the lesion. The second surgical technique matching engine is used to characterize the mapping relationship between the output value of the third surgical technique function and the resection procedure.

[0007] In one embodiment, the processor is further configured to perform the following steps: when there are multiple lesions, the second secondary node is executed to calculate the output value of the third surgical procedure function, the output value of the third surgical procedure function being related to the location of the lesions; when the lesion locations of the multiple lesions are located in the pancreatic head and / or the pancreatic neck, a pancreatic head and body resection auxiliary decision is generated according to the second surgical procedure matching engine; when the lesion locations of the multiple lesions are located in the pancreatic body and / or the pancreatic tail, a pancreatic body and tail resection auxiliary decision is generated according to the second surgical procedure matching engine; when the lesion location is at least one of the pancreatic head and the pancreatic neck, and the lesion location is also at least one of the pancreatic body and the pancreatic tail, a total pancreatectomy auxiliary decision is generated according to the second surgical procedure matching engine.

[0008] In one embodiment, the processor is further configured to perform the following steps: when the number of lesions is one, the first secondary node is executed to determine the nature of the lesion; when the nature of the lesion is the first lesion type, the first tertiary node is executed to calculate the output value of the first surgical procedure function; the first surgical procedure function is constructed based on the lesion parameters; the first surgical procedure matching engine includes a first sub-engine; and the resection surgical procedure auxiliary decision is generated based on the first sub-engine; the lesion parameters include at least one of the lesion location and the distance between the lesion and the blood vessel; the lesion location includes the lesion being located in the pancreatic head, the lesion being located in the pancreatic neck, the lesion being located in the pancreatic body, or the lesion being located in the pancreatic tail; the distance between the lesion and the blood vessel includes at least one of the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein; the first surgical procedure matching sub-engine is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection surgical procedure.

[0009] In one embodiment, the processor is further configured to perform the following steps: when the lesion location is either in the pancreatic head or the pancreatic neck, the lesion parameter is the lesion location, and the output value of the first surgical procedure function is related to the lesion location; when the lesion location is in the pancreatic head, a pancreaticoduodenectomy auxiliary decision is generated based on the first sub-engine; when the lesion location is in the pancreatic neck, a combined pancreatic head and body resection is generated based on the first sub-engine.

[0010] In one embodiment, the processor is further configured to perform the following steps: when the lesion location is in the pancreatic body or the pancreatic tail, the lesion parameters include the lesion location and the distance between the lesion and the blood vessel; when each of the following distances is greater than a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, a pancreatic body and tail resection auxiliary decision is generated according to the first sub-engine; when the lesion location is in the pancreatic body or the pancreatic tail, the lesion parameters include the lesion location and the distance between the lesion and the blood vessel; when any one of the following distances is less than or equal to a first threshold: the pancreatic body and tail extended resection auxiliary decision is generated according to the first sub-engine.

[0011] In one embodiment, the processor is further configured to perform the following steps: when there is only one lesion, the first secondary node is executed to determine the nature of the lesion; when the nature of the lesion is the second lesion type, the second tertiary node is executed to calculate the output value of the second surgical procedure function; the second surgical procedure function is constructed based on the lesion parameters; the first surgical procedure matching engine includes a second sub-engine; the resection surgical procedure auxiliary decision is generated based on the second sub-engine; the lesion parameters include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel; the lesion location includes the lesion being located in the pancreatic head, the lesion being located in the pancreatic neck, the lesion being located in the pancreatic body, or the lesion being located in the pancreatic tail; the distance between the lesion and a blood vessel includes at least one of the following: distance between the lesion and the splenic artery, distance between the lesion and the superior mesenteric artery, distance between the lesion and the splenic vein, distance between the lesion and the superior mesenteric vein, and distance between the lesion and the portal vein; the second sub-engine is used to characterize the mapping relationship between the output value of the second surgical procedure function and the resection surgical procedure.

[0012] In one embodiment, the processor is further configured to perform the following steps: when the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct are both greater than a second threshold, the lesion parameters are the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct, and a local resection auxiliary decision is generated based on the second sub-engine;

[0013] When the lesion location is the pancreatic head, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, and the distance between the lesion and the bile duct; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, the second sub-engine generates an auxiliary decision for pancreatic head resection with preservation of the duodenum.

[0014] When the lesion location is the pancreatic neck, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to a second threshold, and when each of the following distances is greater than a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, a decision-making aid for mid-pancreatic resection with preservation of the duodenum is generated according to the second sub-engine;

[0015] When the lesion is located in the pancreatic neck, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to a second threshold, and when the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, or the distance between the lesion and the portal vein is less than or equal to a first threshold, an auxiliary decision for combined pancreatic head and body resection is generated based on the second sub-engine;

[0016] When the lesion location is either in the pancreatic body or the pancreatic tail, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, and the distance between the lesion and the bile duct; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, a spleen-preserving pancreatic body and tail resection auxiliary decision is generated based on the second sub-engine.

[0017] In one embodiment, the location of the lesion is determined by the processor performing the following steps: based on the three-dimensional medical image of the pancreas, the pancreas in the three-dimensional medical image of the pancreas is extracted into a skeleton to obtain the skeleton center point of the pancreas, and the skeleton center point is sorted according to a first preset rule to form a skeleton center point set; the skeleton center point set is divided into a first subset, a second subset, a third subset, and a fourth subset according to a second preset rule;

[0018] Obtain the centroid of the lesion, calculate the distance between the centroid of the lesion and each skeleton center point in the set of skeleton center points, and determine the skeleton center point closest to the centroid of the lesion as the center point of the lesion location.

[0019] When the center point of the lesion is located in the first subset, the lesion is located in the head of the pancreas; when the center point of the lesion is located in the second subset, the lesion is located in the neck of the pancreas; when the center point of the lesion is located in the third subset, the lesion is located in the body of the pancreas; when the center point of the lesion is located in the fourth subset, the lesion is located in the body of the pancreas.

[0020] One embodiment of this specification provides a method for assisting decision-making in pancreatic lesion resection. The method includes: acquiring three-dimensional medical images of the pancreas, the three-dimensional medical images of the pancreas including lesions; identifying the number and nature of lesions, the number of lesions being one or more, the nature of the lesions including at least one of a first lesion type and a second lesion type; and constructing a decision tree model based on the number of lesions, the nature of the lesions, and a surgical procedure function to generate an assisting decision on the resection procedure corresponding to the lesion.

[0021] One embodiment of this specification provides a pancreatic lesion resection surgical procedure auxiliary decision-making device. The device includes: a medical image acquisition module for acquiring three-dimensional medical images of the pancreas, the three-dimensional medical images of the pancreas including lesions; a lesion identification module for identifying the number and nature of lesions, the number of lesions being one or more, the lesion nature including at least one of a first lesion type and a second lesion type; and an auxiliary decision-making module for constructing a decision tree model based on the number of lesions, the lesion nature, and a surgical procedure function to generate an auxiliary decision for the resection surgical procedure corresponding to the lesion.

[0022] One embodiment of this specification provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring a three-dimensional medical image of the pancreas, the three-dimensional medical image of the pancreas including lesions; identifying the number and nature of the lesions, the number of lesions being one or more, the nature of the lesions including at least one of a first lesion type and a second lesion type; constructing a decision tree model based on the number of lesions, the nature of the lesions, and a surgical procedure function to generate an auxiliary decision for the resection procedure corresponding to the lesions.

[0023] One embodiment of this specification provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program performs the following steps: acquiring a three-dimensional medical image of the pancreas, the three-dimensional medical image of the pancreas including lesions; identifying the number and nature of the lesions, the number of lesions being one or more, the nature of the lesions including at least one of a first lesion type and a second lesion type; constructing a decision tree model based on the number of lesions, the nature of the lesions, and a surgical procedure function to generate an auxiliary decision for the resection procedure corresponding to the lesions.

[0024] Compared with the prior art, the present invention provides a decision-making support system, method, device, computer equipment, and storage medium for pancreatic lesion resection, which has the following beneficial effects:

[0025] Taking into full account the specificity of different pancreatic lesions, a decision tree model for lesion resection is constructed based on the number of lesions, lesion nature, and surgical procedure function. According to the specific situation of different lesions, the decision tree model is used to generate corresponding resection procedures to assist in decision-making, accurately, efficiently, and intelligently assisting doctors in selecting the resection procedure for pancreatic lesions. Attached Figure Description

[0026] Figure 1 A schematic diagram of a pancreatic lesion resection surgical procedure auxiliary decision-making system 100 in one embodiment;

[0027] Figure 2 This is a schematic diagram of the execution flow of a pancreatic lesion resection surgical auxiliary decision-making system in one embodiment;

[0028] Figure 3 This is a schematic diagram of an auxiliary decision tree model for pancreatic lesion resection in one embodiment;

[0029] Figure 4 This is a structural block diagram of a pancreatic lesion resection surgical procedure auxiliary decision-making device in one embodiment;

[0030] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0032] Figure 1 This is a schematic diagram of a pancreatic lesion resection procedure auxiliary decision-making system 100 according to an embodiment of this application. The pancreatic lesion resection procedure auxiliary decision-making system 100 may include a data acquisition device 110, a processing device 120, a storage device 140, and an interaction device 150. The data acquisition device 110, processing device 120, storage device 140, and interaction device 150 can communicate with each other via a network 130. The processing device 120 includes a processor.

[0033] The data acquisition device 110 can be a device for acquiring data. The acquired data may include image data, object feature data, pancreatic medical images in different modalities, pancreatic medical images at different phases, or pancreatic medical images at different time sequences, etc. For example, the acquired image data may be a three-dimensional medical image of the pancreas. In one embodiment, the data acquisition device 110 may include one or more imaging devices. The imaging device can acquire image data. The imaging device may be one or more combinations of magnetic resonance imaging (MRI), computed tomography (CT), positron emission tomography (PET), b-scan ultrasonography, diasonography, thermal texture maps (TTM), and medical electronic endoscope (MEE). The image data may be a three-dimensional medical image of the pancreas. The image data may be images or data including blood vessels, tissues, or organs of the object. Image data may include multi-phase enhanced CT images, which may include arterial phase, portal venous phase, and delayed phase. Image data may also include time-series MRI images, including T1-weighted images, T2-weighted images, diffusion-weighted images, dynamic contrast-enhanced MRI images, and MRI pancreaticobiliary images. In one embodiment, an object feature acquisition device may be integrated into the imaging device to simultaneously acquire image data and object feature data. In one embodiment, the data acquisition device 110 may transmit the acquired data to a processing device 120, a storage device 140, and / or an interactive device 150, etc., via a network 130.

[0034] The processing device 120 can process data. This data may be data acquired by the data acquisition device 110, data read from the storage device 140, feedback data obtained from the interaction device 150 (such as user input data), or data obtained from the cloud or external devices via the network 130. In one embodiment, the data may include image data, object feature data, user input data, decision tree model data, etc. Data processing may include selecting a region of interest (ROI) in the image data. The ROI may be selected automatically by the processing device 120 or based on user input data. In one embodiment, the selected ROI may be a lesion, blood vessel, tissue, or organ. For example, the ROI may be a pancreatic organ, lesion, pancreatic duct, bile duct, splenic artery, superior mesenteric artery, splenic vein, superior mesenteric vein, or portal vein. The processing device 120 may further segment the ROI in the image. Image segmentation methods can include edge-based methods such as the Perwitt operator, Sobel operator, gradient operator, and Kirch operator; region-based methods such as region growing, thresholding, and clustering; and other segmentation methods such as those based on fuzzy sets and neural networks.

[0035] In one embodiment, the processing device 120 can register medical images across multiple phases. For example, it receives CT arterial phase, CT portal venous phase, and CT delayed phase medical images acquired by the data acquisition device 110, and registers these CT images across multiple phases based on image features to obtain a first modality of pancreatic medical images. It can also receive MRI T1-weighted images, T2-weighted images, diffusion-weighted images, dynamic contrast-enhanced MRI images, and MRI pancreaticobiliary images acquired by the data acquisition device 110, and register these MRI images across multiple phases based on image features to obtain a second modality of pancreatic medical images. The first modality of pancreatic medical images and the second modality of pancreatic medical images are then registered to obtain a registered pancreatic medical image. In one embodiment, the registration method includes rigid registration, elastic registration, etc.

[0036] In one embodiment, the processing device 120 can automatically identify the benign or malignant nature of pancreatic tumors based on deep learning classification algorithms such as the classic VGG16, through steps such as data annotation and model training, and output the number of lesions and lesion nature labels. In one embodiment, the lesion nature includes at least one of a first lesion type and a second lesion type. The first lesion type includes malignant tumors, and the second lesion type includes benign tumors and / or low-grade malignant tumors.

[0037] In one embodiment, the processing device 120 can construct a decision tree model based on the number of lesions, lesion characteristics, and surgical procedure function to generate an auxiliary decision for the resection procedure corresponding to each lesion. In another embodiment, the processing device 120 can determine lesion parameters, including at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel. The processing device 120 can construct a surgical procedure function based on the lesion parameters and generate an auxiliary decision for the resection procedure corresponding to each lesion based on the output value of the surgical procedure function.

[0038] In one embodiment, the processing device 120 may perform noise reduction or smoothing on the acquired data or processing results. In another embodiment, the processing device 120 may send the acquired data or processing results to the storage device 140 for storage or to the interactive device 150 for display. The processing result may be an intermediate result generated during the processing, such as identifying the number and nature of lesions, or it may be the final result of the processing, such as generating a surgical procedure to assist in decision-making for the lesion. In one embodiment, the processing device 120 may be one or more processing elements or devices, such as a central processing unit (CPU), graphics processing unit (GPU), digital signal processor (DSP), system on a chip (SoC), microcontroller unit (MCU), etc. In another embodiment, the processing device 120 may also be a specially designed processing element or device with special functions. The processing device 120 may be local or remote relative to the data acquisition device 110.

[0039] Storage device 140 can store data or information. The data or information may include data acquired by data acquisition device 110, processing results or control commands generated by processing device 120, and user input data received by interactive device 150. Storage device 140 can be one or more read- or write-able storage media, including static random access memory (SRAM), random-access memory (RAM), read-only memory (ROM), hard disk, flash memory, etc. In one embodiment, storage device 140 can also be a remote storage device, such as a cloud drive.

[0040] The interactive device 150 can receive, send, and / or display data or information. The received data or information may include data acquired by the data acquisition device 110, processing results generated by the processing device 120, and data stored by the storage device 140. For example, the data or information displayed by the interactive device 150 may include three-dimensional medical images of the pancreas obtained by the data acquisition device 110, the number and nature of lesions segmented and identified by the processing device 120, the surgical procedure function calculated by the processing device 120, the decision tree model constructed by the processing device 120, and the lesion resection procedure auxiliary decision calculated by the processing device 120. The display format may include one or more combinations of two-dimensional color medical images, three-dimensional color medical images, color geometric models and their mesh analysis, vector graphics (such as velocity vector lines), contour maps, filled contour maps (cloud maps), XY scatter plots, particle trajectory maps, and simulated flow effects. As another example, the data or information sent by the interactive device 150 may include user input information. The interactive device 150 can receive one or more operating parameters input by the user and send them to the processing device 120. In one embodiment, the interactive device 150 may include a user interface. A user can input data into the interactive device 150 using a specific interactive device, such as a mouse, keyboard, touchpad, microphone, etc.

[0041] In one embodiment, the interactive device 150 may be a display screen or other device with display functionality. In one embodiment, the interactive device 150 may have some or all of the functions of the processing device 120. For example, the interactive device 150 may perform operations such as smoothing, noise reduction, and color changing on the results generated by the processing device 120. For example, a color changing operation can convert a grayscale image into a color image, or a color image into a grayscale image. In one embodiment, the interactive device 150 and the processing device 120 may be an integrated device. The integrated device can simultaneously implement the functions of the processing device 120 and the interactive device 150. In one embodiment, the interactive device 150 may include a desktop computer 150-2, a server, a mobile device, etc. Mobile devices may include mobile phones 150-1, laptops, tablets 150-3, built-in devices in vehicles (e.g., motor vehicles, ships, airplanes, etc.), wearable devices, etc. In one embodiment, the interactive device 150 may include or be connected to a display device, printer, fax machine, etc.

[0042] Network 130 can be used for communication within the pancreatic lesion resection surgical auxiliary decision-making system 100, receiving information from outside the system, and sending information to outside the system. In one embodiment, the data acquisition device 110, processing device 120, and interaction device 150 can access network 130 via wired connection, wireless connection, or a combination thereof. Network 130 can be a single network or a combination of multiple networks. In one embodiment, network 130 can be one or more of, but not limited to, local area networks (LANs), wide area networks (WANs), public networks, private networks, wireless LANs, virtual networks, metropolitan area networks (MANs), and public switch telephone networks. In one embodiment, network 130 can include multiple network access points, such as wired or wireless access points, base stations, or network switching points, through which data sources connect to network 130 and send information via the network.

[0043] In one embodiment, such as Figure 2 The flowchart 200 provided illustrates an execution flow for a pancreatic lesion resection surgical auxiliary decision-making system. The system includes a processor, which performs the following steps:

[0044] Step 210: Obtain three-dimensional medical images of the pancreas, including lesions.

[0045] In one embodiment, the three-dimensional medical image of the pancreas may include XR (X-ray) images, CT images, MRI images, ultrasound images, or any combination thereof. The three-dimensional medical image of the pancreas includes lesions. Lesions may be malignant tumors, benign tumors, or low-grade malignant tumors.

[0046] Step 220: Identify the number and nature of lesions. The number of lesions may be one or more, and the nature of the lesions may include at least one of the first lesion type and the second lesion type.

[0047] In one embodiment, the processor, based on deep learning classification algorithms such as VGG16 (Visual GeometryGroup 16-layer Convolutional Neural Network), nnUNet (Normalized U-Net, a normalized U-shaped network that does not require manual parameter tuning), and SWIN-Unet (SwinTransformer-based U-Net, a U-shaped network based on a sliding window attention mechanism), automatically identifies benign and malignant pancreatic tumors through data annotation and model training, outputting the number of lesions and lesion nature labels, thereby improving diagnostic accuracy. In one embodiment, the lesion nature includes at least one of a first lesion type and a second lesion type. The first lesion type includes malignant tumors, and the second lesion type includes benign tumors and / or low-grade malignant tumors.

[0048] Step 230: Construct a decision tree model based on the number of lesions, the nature of the lesions, and the surgical procedure function to generate auxiliary decision-making for the resection procedure corresponding to the lesions.

[0049] In one embodiment, the number of lesions may be one or more belonging to different decision tree branches. The first lesion type or the second lesion type may belong to different decision tree branches. Different output values ​​of the surgical procedure function are used to characterize different resection surgical procedure decisions. The resection surgical procedure decision is used to provide decision suggestions to the physician; the specific diagnosis and treatment plan are subject to the physician's judgment.

[0050] The above method fully considers the specificity of different pancreatic lesions. Based on the number of lesions, the nature of the lesions, and the surgical procedure function, a decision tree model for lesion resection is constructed. According to the specific situation of different lesions, the decision tree model is used to generate corresponding resection procedures to assist in decision-making. This method accurately, efficiently, and intelligently assists doctors in selecting resection procedures for pancreatic lesions, helps doctors to make more precise pancreatic tumor treatment plans, and improves the accuracy of surgery and treatment results.

[0051] Step 210: When the processor acquires three-dimensional medical images of the pancreas, it also includes performing the following steps:

[0052] Step 211: Acquire medical images in multiple modalities, including a first modality and a second modality.

[0053] In one embodiment, the first modality is CT images and the second modality is MRI images.

[0054] Step 212: Acquire multiple phases of medical images in the first mode, and register the multiple phases of medical images acquired in the first mode to obtain the first mode pancreatic medical image.

[0055] In one embodiment, acquiring multiple phases of medical images in the first modality includes CT arterial phase medical images, CT portal venous phase medical images, and CT delayed phase medical images. CT arterial phase medical images show the blood supply to pancreatic lesions, CT portal venous phase medical images observe the portal venous system and liver metastases, and CT delayed phase medical images assess fibrotic or cystic components. Using the portal venous phase as a baseline, arterial and delayed phase data are registered to achieve registration between multiple phases, resulting in the first modality of pancreatic medical images, i.e., multi-phase registered CT modality of pancreatic medical images. In one embodiment, respiratory gating or breath-hold scanning is used to reduce respiratory artifacts when acquiring multiple phases of medical images in the first modality.

[0056] Step 213: Acquire multiple time-series medical images in the second modality, and register the multiple time-series medical images acquired in the second modality to obtain the second modality pancreatic medical image.

[0057] In one embodiment, multiple time-series medical images are acquired in the second modality, including T1-weighted images, T2-weighted images, diffusion-weighted images, dynamic contrast-enhanced MRI images, and MRI pancreaticobiliary images. T1-weighted or T2-weighted images are used for contrast between the pancreatic parenchyma and surrounding fat; diffusion-weighted images are used for tumor cell density assessment; dynamic contrast-enhanced MRI images are used to enhance the visualization of pancreatic tissue; and MRI pancreaticobiliary images are used to visualize pancreatic duct or bile duct dilation. Deformation at each phase is captured using optical flow methods, and registration between multiple time-series images is achieved to obtain the second modality of pancreatic medical images, i.e., multi-time-series registered MRI modality of pancreatic medical images. In one embodiment, navigation echo triggering is used to reduce respiratory artifacts when acquiring multiple time-series medical images in the second modality.

[0058] Step 214: Register the first modality of pancreatic medical images and the second modality of pancreatic medical images to obtain the registered pancreatic medical images.

[0059] In one embodiment, using the pancreas as the region of interest, rigid registration is performed on a first modality of pancreatic medical images and a second modality of pancreatic medical images. Then, a symmetric normalization algorithm is used to optimize local alignment, followed by elastic registration of the rigidly registered first and second modality pancreatic medical images. In another embodiment, vascular-guided registration is used, extracting the celiac trunk and portal vein as registration references to enhance vascular alignment accuracy. In one embodiment, the registered pancreatic medical images are located in CT image space. In yet another embodiment, the registered pancreatic medical images are located in MRI image space.

[0060] Step 215: Perform three-dimensional reconstruction on the registered pancreatic medical images to obtain three-dimensional pancreatic medical images.

[0061] In one embodiment, 3D reconstruction of the registered pancreatic medical image includes segmentation of tissue and organ structures. This segmentation includes the segmentation of the pancreas, lesions, pancreatic ducts, bile ducts, blood vessels, and surrounding organs. Blood vessels include the splenic artery, superior mesenteric artery, splenic vein, superior mesenteric vein, and portal vein. Surrounding organs include the stomach, duodenum, or spleen. A pancreatic segmentation model pre-trained using the nnUNet algorithm can be used for pancreatic segmentation. The SWIN-Unet algorithm can be used for lesion segmentation. U-Net (U-shaped network) combined with a blood vessel tracking algorithm can be used to extract tubular structures for pancreatic and bile duct segmentation. Vesselness filtering (vascular enhancement filtering) combined with region growing can be used for blood vessel segmentation. Deep learning methods such as V-Net (V-shaped network) can also be used for blood vessel segmentation. Automatic whole-body medical image segmentation tools can be used for the segmentation of surrounding organs.

[0062] In one embodiment, three-dimensional reconstruction of the registered pancreatic medical images includes tissue and organ structure fusion. Tissue and organ structure fusion includes vascular fusion, pancreatic duct fusion, bile duct fusion, and multi-phase, multi-temporal lesion fusion. Vascular fusion primarily utilizes CTA (CT Angiography) and / or CTV (CT Venography), supplemented by MRI vessel wall information. Pancreatic duct and bile duct fusion integrate the segmentation results of the pancreatic and bile ducts into the registered pancreatic medical images. Multi-phase, multi-temporal lesion fusion employs a region-weighted fusion approach, for example, prioritizing arterial phase images, increasing the weight of arterial phase images, and decreasing the weight of images from other phases or temporal sequences.

[0063] In one embodiment, 3D reconstruction of the registered pancreatic medical image is performed, including surface reconstruction, volume rendering fusion, and a spatiotemporal dynamic model. Surface reconstruction involves generating smooth surfaces of the pancreas, lesion, and surrounding organs using Marching Cubes. Tubular meshes are generated by extracting vascular centerlines. Volume rendering fusion includes multimodal transparency adjustment, color coding, etc. The spatiotemporal dynamic model includes 4D blood flow simulation to simulate the perfusion dynamics of contrast agent within the lesion.

[0064] The system integrates the advantages of multimodal and multi-temporal imaging, such as CT to show calcification and blood vessels, and MRI to define tumor boundaries. By accurately calculating the positional relationship between the tumor and the pancreatic duct, bile duct and surrounding blood vessels, the system improves the accuracy of assessment and provides an accurate medical imaging basis for subsequent decision-making on the surgical procedure for pancreatic lesion resection. At the same time, it provides users with intuitive display support and reduces reliance on doctors' experience.

[0065] refer to Figure 3 The decision tree model for pancreatic lesion resection is used. The processor executes step 230, which involves constructing a decision tree model based on the number of lesions, lesion characteristics, and surgical procedure function to generate the corresponding resection surgical procedure for each lesion. This also includes the following steps:

[0066] Step 310: Construct a surgical procedure function based on the lesion parameters. The lesion parameters include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and the blood vessel. The surgical procedure function includes a first surgical procedure function, a second surgical procedure function, and a third surgical procedure function.

[0067] In one embodiment, a surgical procedure function is used to characterize the relationship between lesion parameters and lesion resection procedures. Different lesions have different lesion parameters, and different lesion parameters result in different applicable resection procedures. The surgical procedure function is constructed from lesion parameters, which include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel. The surgical procedure function includes a first surgical procedure function y1, a second surgical procedure function y2, and a third surgical procedure function y3.

[0068] Step 320: Construct the root node of the decision tree model based on the number of lesions. When there is only one lesion, construct the first and second level nodes of the decision tree model based on the nature of the lesion. Construct the first and third level nodes of the decision tree model based on the first surgical procedure function, or construct the second and third level nodes of the decision tree model based on the second surgical procedure function. Generate the resection procedure auxiliary decision based on the first surgical procedure matching engine. The first surgical procedure matching engine is used to characterize the mapping relationship between the output value of the first surgical procedure function or the output value of the second surgical procedure function and the resection procedure.

[0069] In one embodiment, the lesion resection procedure is also related to the number and nature of the lesions. The number of lesions serves as the root node of the decision tree model, and the lesion nature serves as the first and second-level nodes. When the lesion is a single lesion, the lesion nature needs to be further considered. Based on the lesion nature, the decision is made to enter either the first or second-level node constructed by the first procedure function y1, or the second or third-level node constructed by the second procedure function y2. The resection procedure auxiliary decision is generated based on the first procedure matching engine, which represents the mapping relationship between the output value of y1 or y2 and the resection procedure.

[0070] Step 330: Construct the root node of the decision tree model based on the number of lesions. When there are multiple lesions, construct the second and second level nodes of the decision tree model based on the third surgical procedure function. Generate the resection surgical procedure auxiliary decision based on the second surgical procedure matching engine. The second surgical procedure matching engine is used to characterize the mapping relationship between the output value of the third surgical procedure function and the resection surgical procedure.

[0071] In one embodiment, the lesion resection technique is also related to the number of lesions. The number of lesions serves as the root node of the decision tree model. When there are multiple lesions, i.e., multiple lesions, the process enters the second-level node of the decision tree model constructed by the third technique function y3. The second technique matching engine generates the resection technique auxiliary decision based on the lesion. The second technique matching engine represents the mapping relationship between the output value of y3 and the resection technique.

[0072] Through the above system, based on the lesion reference system and the lesion function, and based on the decision tree model, the number of lesions, lesion nature, and lesion function are rigorously designed to form an intelligent and efficient pancreatic lesion resection surgical procedure auxiliary decision system. It is applicable to various tumor scenarios such as single, multiple, malignant and benign tumors, and has a wider range of applications, helping users to accurately and quickly determine the resection procedure for specific lesions.

[0073] The location of the lesion in step 310 is determined by the processor performing the following steps:

[0074] Step 311: Based on the three-dimensional medical image of the pancreas, the pancreas in the three-dimensional medical image of the pancreas is extracted into a skeleton to obtain the center point of the pancreas skeleton. The center point of the skeleton is sorted according to the first preset rule to form a set of center points of the skeleton. The set of center points of the skeleton is divided into a first subset, a second subset, a third subset, and a fourth subset according to the second preset rule.

[0075] In one embodiment, a U-Net variant combined with skeleton regression, topology refinement, distance transformation, or end-to-end centerline prediction can be used to extract the skeleton of the pancreas from a 3D medical image, obtaining the pancreatic skeleton center points. In one embodiment, the left side of the 3D medical image represents the pancreatic head, and the right side represents the pancreatic tail. The first preset rule is to sort the skeleton center points from left to right, forming a skeleton center point set. The second preset rule is to divide the skeleton center point set from left to right, with the first 30% (inclusive) of the skeleton center point set as the first subset, 30% (exclusive) to 40% (inclusive) as the second subset, 40% (exclusive) to 75% (inclusive) as the third subset, and 75% (exclusive) to 100% (inclusive) as the fourth subset. Of course, if the patient is a mirror image, the first preset rule can also be to sort the skeleton center points from right to left, forming the skeleton center point set. The second preset rule is to divide the skeleton center point set from right to left, and to divide the subsets from right to left with reference to the subset division ratio from left to right, which will not be elaborated further.

[0076] Step 312: Obtain the centroid of the lesion, calculate the distance between the centroid of the lesion and each skeletal center point in the set of skeletal center points, and determine the skeletal center point closest to the centroid of the lesion as the center point of the lesion location.

[0077] In one embodiment, the distance between the centroid of the lesion and each skeletal center point in the set of skeletal center points is calculated one by one using a distance calculation method until the skeletal center point closest to the centroid of the lesion is found as the lesion location center point. The lesion location center point is used to represent the location of the lesion. In one embodiment, the centroid of the lesion can be determined by spatial moment calculation, keypoint regression, heatmap regression, principal component analysis, or centroid calculation. In one embodiment, the distance calculation method can be Euclidean distance, etc.

[0078] Step 313: When the center point of the lesion is located in the first subset, the lesion is located in the head of the pancreas; when the center point of the lesion is located in the second subset, the lesion is located in the neck of the pancreas; when the center point of the lesion is located in the third subset, the lesion is located in the body of the pancreas; when the center point of the lesion is located in the fourth subset, the lesion is located in the body of the pancreas.

[0079] In one embodiment, different subsets of the center point of the lesion location represent different lesion locations.

[0080] The above system can accurately determine the location of lesions, providing accurate lesion parameter support for subsequent surgical decisions on pancreatic lesion resection.

[0081] Step 330 also includes the processor performing the following steps:

[0082] Step 331: When there are multiple lesions, the second and second-level nodes are executed to calculate the output value of the third surgical function. The output value of the third surgical function is related to the location of the lesion.

[0083] Step 332: When the lesion is located in the pancreatic head or the pancreatic neck, the second surgical procedure matching engine generates an auxiliary decision for combined pancreatic head and body resection.

[0084] Step 333: When the lesion location is either in the pancreatic body or in the pancreatic tail, generate a pancreatic body and tail resection auxiliary decision based on the second surgical procedure matching engine.

[0085] Step 334: When the lesion location is at least one of the lesion in the pancreatic head and the lesion in the pancreatic neck, and the lesion location is also at least one of the lesion in the pancreatic body and the lesion in the pancreatic tail, a total pancreatectomy auxiliary decision is generated according to the second surgical procedure matching engine.

[0086] In one embodiment, the third technique function is The output value of the third surgical function is related to the location of the lesion, and y3 is expressed by the following formula (1).

[0087] (1)

[0088] When the number of lesions is multiple, the processor executes the calculation of the third surgical procedure function, where a i is used to represent the lesion location of the i-th lesion, i = (1, 2, 3, ……, N), δ(a i , 1) and δ(a i , 3) are respectively used to represent the Kronecker function, and N is used to represent the number of lesions. When a i = 1, δ(a i , 1) = 1; when a i ≠ 1, δ(a i , 1) = 0; when a i = 3, δ(a i , 3) = ①; when a i ≠ 3, δ(a i , 3) = 0. a i is assigned different values according to different lesion locations, the output value of y3 is different, and the corresponding assisted decision for the resection procedure is different. The mapping relationship between the output value of the third surgical procedure function represented by the second surgical procedure matching engine and the resection procedure is specifically referred to Table 1.

[0089] Referring to Table 1, when the lesion location is that the lesion is located in the pancreatic head, a i = 1, when the lesion is located in the pancreatic neck, a i = 2, when the N lesion locations are all in the pancreatic head, the function value of y3 is 2; when the N lesion locations are all in the pancreatic neck, the function value of y3 is 2; when some of the N lesions are located in the pancreatic head and some are located in the pancreatic neck, the function value of y3 is 2. According to the second surgical procedure matching engine, y3 = 2, and an assisted decision for a combined resection of the pancreatic head and body is generated for the current multiple lesions.

[0090] Referring to Table 1, when the lesion location is that the lesion is located in the pancreatic body, a i = 3, when the lesion is located in the pancreatic tail, a i = 4, when the N lesion locations are all in the pancreatic body, the function value of y3 is 4; when the N lesion locations are all in the pancreatic tail, the function value of y3 is 4; when some of the N lesions are located in the pancreatic body and some are located in the pancreatic tail, the function value of y3 is 4. According to the second surgical procedure matching engine, y3 = 4, and an assisted decision for a resection of the pancreatic body and tail is generated for the current multiple lesions.<>

[0091] Referring to Table 1, at least one of the N lesions has a lesion location of being located in the pancreatic head or being located in the pancreatic neck, and at least one of the N lesions also has a lesion location of being located in the pancreatic body or being located in the pancreatic tail. According to the calculation of formula (1), it can be known that 2 < y3 < 4. According to the second surgical procedure matching engine, 2 < y3 < 4, and an assisted decision for a total pancreatectomy is generated for the current multiple lesions.

[0092] Note: There seems to be a mistake in the original text where "δ(a i , 3) = ①" should probably be "δ(a i , 3) = 1". This has been corrected in the translation.With the above system, when there are multiple lesions, a surgical procedure function can be constructed based solely on the location of the lesions. A decision tree model can be constructed based on the number of lesions and the third surgical procedure function. Based on the output value of the third surgical procedure function, the surgical procedure for pancreatic lesion resection can be determined quickly, accurately, and efficiently.

[0093] Table 1. Second Technique Matching Engine

[0094]

[0095] Step 320 also includes the processor performing the following steps:

[0096] Step 3210: When there is only one lesion, execute the first and second level nodes to determine the nature of the lesion. When the nature of the lesion is the first lesion type, execute the first and third level nodes to calculate the output value of the first surgical function.

[0097] Step 3211: Construct a first surgical procedure function based on lesion parameters. The first surgical procedure matching engine includes a first sub-engine. The first sub-engine generates a resection surgical procedure auxiliary decision based on the lesion. The lesion parameters include at least one of the lesion location and the distance between the lesion and the blood vessel. The lesion location includes the lesion being located in the pancreatic head, pancreatic neck, pancreatic body, or pancreatic tail. The distance between the lesion and the blood vessel includes at least one of the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein. The first sub-engine is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection surgical procedure.

[0098] In one embodiment, the first surgical function is y1, which is associated with at least one of the lesion location and the distance between the lesion and the blood vessel. The first surgical function y1 is expressed by the following formula (2).

[0099] (2)

[0100] When there is only one lesion, the processor further determines the nature of the lesion. If the lesion is classified as the first type, i.e., malignant tumor, then the first surgical function is calculated, where 'a' represents the location of the lesion. 'a' is assigned a value of 1 when the lesion is in the pancreatic head, 2 when it is in the pancreatic neck, 3 when it is in the pancreatic body, and 4 when it is in the pancreatic tail. δ(a,4) represents the Kronecker function; δ(a,4) = 1 when the lesion is in the pancreatic tail, and 0 when the lesion is in other locations. To round down, for example, if the lesion is located in the pancreatic tail, a=4. =1; if the lesion is located in the pancreatic head, a=1. =0. c1 represents the distance between the lesion and the splenic artery, c2 represents the distance between the lesion and the superior mesenteric artery, c3 represents the distance between the lesion and the splenic vein, c4 represents the distance between the lesion and the superior mesenteric vein, and c5 represents the distance between the lesion and the portal vein.

[0101] Using the above system, when there is only one lesion, a decision tree model is constructed based on the number of lesions, the nature of the lesion, and the first surgical procedure function. The first sub-engine, as shown in Table 2, is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection procedure. Based on the output value of the first surgical procedure function, the resection procedure for pancreatic lesions can be determined quickly, accurately, and efficiently to assist in decision-making.

[0102] Step 3211 also includes the processor performing the following steps:

[0103] Step 32110: When the lesion is located in the head of the pancreas or in the neck of the pancreas, the lesion parameter is the lesion location, and the output value of the first surgical function is related to the lesion location.

[0104] Step 32111: When the lesion is located in the head of the pancreas, generate a pancreaticoduodenectomy auxiliary decision based on the first sub-engine.

[0105] Step 32112: When the lesion is located in the pancreatic neck, generate a combined pancreatic head and body resection based on the first sub-engine.

[0106] In one embodiment, referring to Table 2, when the lesion is located in the head of the pancreas or in the neck of the pancreas, the first surgical function y1 is only related to the location of the lesion, and formula (2) can be simplified to the following formula (3).

[0107] y1=a-δ(a,4) (3)

[0108] When the lesion is located in the pancreatic head (a=1, δ(a,4)=0, y1=1), referring to the first sub-engine in Table 2, a pancreaticoduodenectomy auxiliary decision is generated. When the lesion is located in the pancreatic neck (a=2, δ(a,4)=0, y1=2), referring to the first sub-engine in Table 2, a combined pancreatic head and body resection is generated. Through this system, when there is only one lesion, a decision tree model is constructed based on the number of lesions, the nature of the lesion, and the first surgical procedure function. The first sub-engine, referring to Table 2, is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection procedure. Based on the output value of the first surgical procedure function, the auxiliary decision for pancreatic lesion resection can be determined quickly, accurately, and efficiently. Furthermore, the first surgical procedure function can be simplified based on the lesion location, making the computational efficiency of the auxiliary decision for pancreatic lesion resection even higher.

[0109] Step 3211 also includes the processor performing the following steps:

[0110] Step 32120: When the lesion is located in the pancreatic body or the pancreatic tail, the lesion parameters include the lesion location and the distance between the lesion and the blood vessels. When each of the following distances is greater than the first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, a pancreatic body and tail resection auxiliary decision is generated based on the first sub-engine.

[0111] Step 32121: When the lesion is located in the pancreatic body or the pancreatic tail, the lesion parameters include the lesion location and the distance between the lesion and the blood vessels. If any one of the following distances is less than or equal to a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, an auxiliary decision for extended pancreatic body and tail resection is generated based on the first sub-engine.

[0112] In one embodiment, referring to Table 2, when the lesion is located in the pancreatic body or the pancreatic tail, the first surgical function is related to the lesion location and the distance between the lesion and the blood vessel. The first surgical function is y1, which is expressed by formula (2). When the distance between the lesion and the splenic artery is greater than the first threshold, c1=1; when it is less than or equal to the first threshold, c1=0. When the distance between the lesion and the superior mesenteric artery is greater than the first threshold, c2=1; when it is less than or equal to the first threshold, c2=0. When the distance between the lesion and the splenic vein is greater than the first threshold, c3=1; when it is less than or equal to the first threshold, c3=0. When the distance between the lesion and the superior mesenteric vein is greater than the first threshold, c4=1; when it is less than or equal to the first threshold, c4=0. When the distance between the lesion and the portal vein is greater than the first threshold, c5=1; when it is less than or equal to the first threshold, c5=0.

[0113] When each of the following distances is greater than the first threshold, the output value y1 of the first surgical procedure function is 3, and the auxiliary decision for pancreatic distalization and distalization is generated according to the first sub-engine in Table 2.

[0114] When any one of the following distances is less than or equal to the first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, the output value y1 of the first surgical procedure function is 4, and the auxiliary decision for extended pancreatic body and tail resection is generated according to the first sub-engine in Table 2.

[0115] In one embodiment, the first threshold is 1 mm. The distance relationships between the lesion and the splenic artery, the lesion and the superior mesenteric artery, the lesion and the splenic vein, the lesion and the superior mesenteric vein, and the lesion and the portal vein can have the same first threshold or different first thresholds. The first threshold can be adjusted according to the user's needs.

[0116] Table 2 First Sub-engine

[0117]

[0118] In Table 2, “ / ” indicates any value that the parameter can take within the range of values: a∈{1,2,3,4}, c1∈{0,1}, c2∈{0,1}, c3∈{0,1}, c4∈{0,1}, c5∈{0,1}.

[0119] Step 320 also includes the processor performing the following steps:

[0120] Step 3220: When there is only one lesion, execute the first and second level nodes to determine the nature of the lesion; when the nature of the lesion is the second lesion type, execute the second and third level nodes to calculate the output value of the second surgical function.

[0121] Step 3221: Construct a second surgical procedure function based on lesion parameters. The first surgical procedure matching engine includes a second sub-engine. The second sub-engine generates a resection surgical procedure auxiliary decision based on the lesion. The lesion parameters include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel. The lesion location includes the lesion being located in the pancreatic head, pancreatic neck, pancreatic body, or pancreatic tail. The distance between the lesion and a blood vessel includes at least one of the following: distance between the lesion and the splenic artery, distance between the lesion and the superior mesenteric artery, distance between the lesion and the splenic vein, distance between the lesion and the superior mesenteric vein, and distance between the lesion and the portal vein. The second sub-engine is used to characterize the mapping relationship between the output value of the second surgical procedure function and the resection surgical procedure.

[0122] In one embodiment, the second surgical function is y2, which is related to at least one of the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel. The second surgical function y2 is expressed by the following formula (4).

[0123] (4)

[0124] When there is only one lesion, the processor further determines the nature of the lesion. If the lesion is classified as a second lesion type, i.e., a benign tumor and / or a low-grade malignant tumor, then the second surgical function is calculated, where 'a' represents the location of the lesion. 'a' is assigned a value of 1 when the lesion is located in the pancreatic head, 2 when the lesion is located in the pancreatic neck, 3 when the lesion is located in the pancreatic body, and 4 when the lesion is located in the pancreatic tail. δ(a,2) δ(a,3) represents the Kronecker function. When the lesion is located in the pancreatic neck, δ(a,2)=1; when the lesion is located elsewhere, δ(a,2)=0. Similarly, when the lesion is located in the pancreatic body, δ(a,3)=1; when the lesion is located elsewhere, δ(a,3)=0. b1 represents the distance between the lesion and the pancreatic duct, b2 represents the distance between the lesion and the bile duct, c1 represents the distance between the lesion and the splenic artery, c2 represents the distance between the lesion and the superior mesenteric artery, c3 represents the distance between the lesion and the splenic vein, c4 represents the distance between the lesion and the superior mesenteric vein, and c5 represents the distance between the lesion and the portal vein. Using this system, when there is only one lesion, a decision tree model is constructed based on the number of lesions, the nature of the lesion, and the second surgical procedure function. The second sub-engine, as shown in Table 3, is used to characterize the mapping relationship between the output value of the second surgical procedure function and the resection procedure. Based on the output value of the second surgical procedure function, the resection procedure for pancreatic lesions can be determined quickly, accurately, and efficiently to assist in decision-making.

[0125] Step 3221 also includes the processor performing the following steps:

[0126] Step 32211: When the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct are both greater than the second threshold, the lesion parameters are the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct, and the local resection auxiliary decision is generated according to the second sub-engine.

[0127] In one embodiment, referring to Table 3, b1=1 when the distance between the lesion and the pancreatic duct is greater than the second threshold, and b1=0 when it is less than or equal to the second threshold; b2=1 when the distance between the lesion and the bile duct is greater than the second threshold, and b2=0 when it is less than or equal to the second threshold. When both the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct are greater than the second threshold, the second surgical function is only related to the distance relationship between the lesion and the bile duct, and formula (4) can be simplified to formula (5).

[0128] (5)

[0129] The output value of the second surgical procedure function is y2=5. Based on the second sub-engine in Table 3, a local resection auxiliary decision is generated.

[0130] Step 32212: When the lesion location is the pancreatic head, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, and the distance between the lesion and the bile duct; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, an auxiliary decision for pancreatic head resection with preservation of the duodenum is generated according to the second sub-engine.

[0131] In one embodiment, referring to Table 3, when the lesion is located in the pancreatic head a=1 and δ(a,3)=0, the second surgical function is related to the location of the lesion, the distance between the lesion and the bile duct, and the distance between the lesion and the bile duct. Formula (4) can be simplified to formula (6).

[0132] (6)

[0133] When the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, b1=0 and / or b2=0, the output value y2 of the second surgical procedure function is 6, and the auxiliary decision for pancreatic head resection with preservation of duodenum is generated according to the second sub-engine in Table 3.

[0134] Step 32213: When the lesion location is the pancreatic neck, and the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to a second threshold, and when each of the following distances is greater than a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, a mid-pancreatic resection with preservation of the duodenum is generated according to the second sub-engine.

[0135] In one embodiment, referring to Table 3, when the lesion is located in the pancreatic neck (a=2), the second surgical procedure function is referenced by formula (4). The second surgical procedure function is related to the location of the lesion, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel. When the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold (b1=0) and / or (b2=0), and the distance between the lesion and the splenic artery is greater than the first threshold (c1=1), the distance between the lesion and the superior mesenteric artery is greater than the first threshold (c2=1), the distance between the lesion and the splenic vein is greater than the first threshold (c3=1), the distance between the lesion and the superior mesenteric vein is greater than the first threshold (c4=1), and the distance between the lesion and the portal vein is greater than the first threshold (c5=1), the output value of the second surgical procedure function is y2=7. Based on Table 3, the second sub-engine generates an auxiliary decision for mid-segment pancreatic resection with preservation of the duodenum.

[0136] Step 32214: When the lesion location is the pancreatic neck, and the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, and when the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, or the distance between the lesion and the portal vein is less than or equal to the first threshold, the auxiliary decision for combined pancreatic head and body resection is generated according to the second sub-engine.

[0137] In one embodiment, referring to Table 3, when the lesion is located in the pancreatic neck (a=2), the second surgical procedure function is referenced by formula (4). The second surgical procedure function is related to the location of the lesion, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel. When the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold (b1=0) and / or (b2=0), and when any one of the following exists (c1=0 when the distance between the lesion and the splenic artery is less than or equal to the first threshold, c2=0 when the distance between the lesion and the superior mesenteric artery is less than or equal to the first threshold, c3=0 when the distance between the lesion and the splenic vein is less than or equal to the first threshold, c4=0 when the distance between the lesion and the superior mesenteric vein is less than or equal to the first threshold, or c5=0 when the distance between the lesion and the portal vein is less than or equal to the first threshold), the output value of the second surgical procedure function is y2=8. The second sub-engine generates the auxiliary decision for pancreatic head and body resection according to Table 3.

[0138] Step 32215: When the lesion location is either in the pancreatic body or the pancreatic tail, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, and the distance between the lesion and the bile duct; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, the spleen-preserving pancreatic body and tail resection auxiliary decision is generated according to the second sub-engine.

[0139] In one embodiment, referring to Table 3, when the lesion location is a=3 in the pancreatic body or a=4 in the pancreatic tail, δ(a,3)=0, the second surgical function is related to the lesion location, the distance between the lesion and the bile duct, and the distance between the lesion and the bile duct. Formula (4) can be simplified to formula (6) above. When the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, b1=0 and / or b2=0, the output value y2 of the second surgical function is 9, and the spleen-preserving pancreatic body and tail resection auxiliary decision is generated according to the second sub-engine in Table 3.

[0140] In one embodiment, the first threshold is 1 mm. The distances between the lesion and the splenic artery, the lesion and the superior mesenteric artery, the lesion and the splenic vein, the lesion and the superior mesenteric vein, and the lesion and the portal vein can all have the same first threshold, or they can have different first thresholds. The first threshold can be adjusted according to the user's needs. In one embodiment, the second threshold is 3 mm. The second threshold used to represent the distance between the lesion and the pancreatic duct can be the same as or different from the second threshold used to represent the distance between the lesion and the bile duct.

[0141] Using the above system, when there is only one lesion, a decision tree model is constructed based on the number of lesions, the nature of the lesion, and the second surgical procedure function. The second sub-engine, as shown in Table 3, represents the mapping relationship between the output value of the second surgical procedure function and the resection procedure. Based on the output value of the second surgical procedure function, the resection procedure for pancreatic lesions can be determined quickly, accurately, and efficiently. Furthermore, the second surgical procedure function can be simplified based on the lesion location, making the computational efficiency of the pancreatic lesion resection procedure auxiliary decision even higher.

[0142] Table 3 Second Sub-engine

[0143]

[0144] In Table 3, “ / ” indicates any value that the parameter can take within the range of values: a∈{1,2,3,4}, b1∈{0,1}, b2∈{0,1}, c1∈{0,1}, c2∈{0,1}, c3∈{0,1}, c4∈{0,1}, c5∈{0,1}.

[0145] It should be understood that, although Figure 2 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 2 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0146] In one embodiment, a method for assisting decision-making regarding pancreatic lesion resection is provided, the method comprising the following steps:

[0147] Acquire three-dimensional medical images of the pancreas, including lesions;

[0148] Identify the number and nature of lesions. The number of lesions may be one or more, and the nature of the lesions may include at least one of the first lesion type and the second lesion type.

[0149] A decision tree model is constructed based on the number of lesions, the nature of the lesions, and the surgical procedure function to generate the corresponding surgical procedure for lesion resection to assist in decision-making.

[0150] In one embodiment, such as Figure 4 As shown, a pancreatic lesion resection surgical procedure auxiliary decision-making device 400 is provided. This device includes a medical image acquisition module 410, a lesion identification module 420, and an auxiliary decision-making module 430, wherein:

[0151] The medical image acquisition module 410 is used to acquire three-dimensional medical images of the pancreas, wherein the three-dimensional medical images of the pancreas include lesions;

[0152] The lesion identification module 420 is used to identify the number and nature of lesions, wherein the number of lesions is one or more, and the nature of the lesions includes at least one of a first lesion type and a second lesion type;

[0153] The auxiliary decision module 430 is used to construct a decision tree model based on the number of lesions, the nature of the lesions, and the surgical procedure function to generate auxiliary decisions for the resection procedure corresponding to the lesions.

[0154] Specific limitations regarding the pancreatic lesion resection surgical decision support device can be found in the above description of the pancreatic lesion resection surgical decision support system, and will not be repeated here. Each module in the aforementioned pancreatic lesion resection surgical decision support device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0155] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for assisting decision-making in pancreatic lesion resection surgery. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0156] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0157] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0158] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0159] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0160] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0161] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A decision-making support system for pancreatic lesion resection, characterized in that, The system includes a processor, the processor being configured to perform the following steps: Acquire three-dimensional medical images of the pancreas, wherein the three-dimensional medical images of the pancreas include lesions; Identify the number and nature of lesions, wherein the number of lesions is one or more, and the nature of the lesions includes at least one of a first lesion type and a second lesion type; A decision tree model is constructed based on the number of lesions, the nature of the lesions, and the surgical procedure function, including: The root node of the decision tree model is constructed based on the number of lesions. When there is only one lesion, the first and second level nodes of the decision tree model are constructed based on the nature of the lesion, and the first and third level nodes of the decision tree model are constructed based on the first formula function, or the second and third level nodes of the decision tree model are constructed based on the second formula function. When there are multiple lesions, the second and second level nodes of the decision tree model are constructed based on the third formula function. The decision tree model is used to generate an auxiliary decision on the surgical procedure for the lesion, including: When there is only one lesion, the first second-level node is executed to determine the nature of the lesion. When the nature of the lesion is the first lesion type, the first third-level node is executed to calculate the output value of the first surgical function. The first surgical procedure function is constructed based on the parameters of the first lesion. The first surgical procedure matching engine includes a first sub-engine, which is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection procedure. The first resection procedure auxiliary decision is generated based on the first sub-engine. The first lesion parameter includes at least one of the lesion location and the distance between the lesion and the blood vessel. The lesion location includes the lesion being located in the pancreatic head, the pancreatic neck, the pancreatic body, or the pancreatic tail. The distance between the lesion and the blood vessel includes at least one of the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein. The first resection procedure auxiliary decision includes: pancreaticoduodenectomy, combined pancreatic head and body resection, pancreatic body and tail resection, and extended pancreatic body and tail resection. When there is only one lesion, the first second-level node is executed to determine the nature of the lesion. When the nature of the lesion is the second lesion type, the second third-level node is executed to calculate the output value of the second surgical function. The second surgical procedure function is constructed based on the second lesion parameters. The first surgical procedure matching engine includes a second sub-engine, which is used to characterize the mapping relationship between the output value of the second surgical procedure function and the resection procedure. The second resection procedure auxiliary decision is generated based on the second sub-engine. The second lesion parameters include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel. The lesion location includes the lesion being located in the pancreatic head, pancreatic neck, pancreatic body, or pancreatic tail. The distance between the lesion and a blood vessel includes at least one of the following: distance between the lesion and the splenic artery, distance between the lesion and the superior mesenteric artery, distance between the lesion and the splenic vein, distance between the lesion and the superior mesenteric vein, and distance between the lesion and the portal vein. The second resection procedure auxiliary decision includes: local resection, pancreatic head resection with preservation of the duodenum, pancreatic head resection with preservation of the duodenum, mid-pancreatic resection with preservation of the duodenum, combined pancreatic head and body resection, and pancreatic body and tail resection with preservation of the spleen. When there are multiple lesions, the second and second-level nodes are executed to calculate the output value of the third surgical procedure function. The output value of the third surgical procedure function is related to the location of the lesion. Based on the output value of the third surgical procedure function, a third resection surgical procedure auxiliary decision is generated for the lesion. The third resection surgical procedure auxiliary decision includes: combined pancreatic head and body resection, pancreatic body and tail resection, and total pancreatectomy.

2. The system according to claim 1, characterized in that, When the processor performs the acquisition of three-dimensional medical images of the pancreas, it also includes performing the following steps: Acquire medical images in multiple modalities, including a first modality and a second modality; Multiple phases of medical images are acquired in the first modality, and the multiple phases of medical images acquired in the first modality are registered to obtain the first modality pancreatic medical image. Multiple time-series medical images are acquired in the second modality, and the multiple time-series medical images acquired in the second modality are registered to obtain the second modality pancreatic medical image; The first modality of pancreatic medical images and the second modality of pancreatic medical images are registered to obtain registered pancreatic medical images; The registered pancreatic medical images are reconstructed in three dimensions to obtain the three-dimensional pancreatic medical images.

3. The system according to claim 1, characterized in that, The processor is also used to perform the following steps: When there are multiple lesions, the second and second-level nodes are executed to calculate the output value of the third surgical function, and the output value of the third surgical function is related to the location of the lesions; When the location of multiple lesions is that the lesion is located in the pancreatic head and / or the lesion is located in the pancreatic neck, the pancreatic head and body combined resection auxiliary decision is generated according to the second surgical procedure matching engine; When the location of multiple lesions is that the lesion is located in the pancreatic body and / or the lesion is located in the pancreatic tail, a pancreatic body and tail resection auxiliary decision is generated according to the second surgical procedure matching engine; When the lesion location is at least one of the lesion located in the pancreatic head and the pancreatic neck, and the lesion location is also at least one of the lesion located in the pancreatic body and the pancreatic tail, a total pancreatectomy auxiliary decision is generated according to the second surgical procedure matching engine; the second surgical procedure matching engine is used to characterize the mapping relationship between the output value of the third surgical procedure function and the resection procedure.

4. The system according to claim 1, characterized in that, The processor is also used to perform the following steps: When the lesion location is either in the pancreatic head or in the pancreatic neck, the lesion parameter is the lesion location, and the output value of the first surgical procedure function is related to the lesion location. When the lesion is located in the head of the pancreas, a pancreaticoduodenectomy auxiliary decision is generated based on the first sub-engine. When the lesion is located in the pancreatic neck, a combined pancreatic head and body resection is generated based on the first sub-engine.

5. The system according to claim 1, characterized in that, The processor is also used to perform the following steps: When the lesion is located in the pancreatic body or the pancreatic tail, the lesion parameters include the lesion location and the distance between the lesion and the blood vessel. When each of the following distances is greater than a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, a pancreatic body and tail resection auxiliary decision is generated based on the first sub-engine. When the lesion is located in the pancreatic body or the pancreatic tail, the lesion parameters include the lesion location and the distance between the lesion and the blood vessel. If any one of the following distances is less than or equal to a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, an auxiliary decision for extended pancreatic body and tail resection is generated based on the first sub-engine.

6. The system according to claim 1, characterized in that, The processor is also used to perform the following steps: When the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct are both greater than the second threshold, the lesion parameters are the distance between the lesion and the pancreatic duct and the distance between the lesion and the bile duct, and a local resection auxiliary decision is generated based on the second sub-engine; When the lesion location is the pancreatic head, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, and the distance between the lesion and the bile duct; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, the second sub-engine generates an auxiliary decision for pancreatic head resection with preservation of the duodenum. When the lesion location is the pancreatic neck, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to a second threshold, and when each of the following distances is greater than a first threshold: the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein, a decision-making aid for mid-pancreatic resection with preservation of the duodenum is generated according to the second sub-engine; When the lesion is located in the pancreatic neck, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, the distance between the lesion and the bile duct, and the distance between the lesion and the blood vessel; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to a second threshold, and when the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, or the distance between the lesion and the portal vein is less than or equal to a first threshold, an auxiliary decision for combined pancreatic head and body resection is generated based on the second sub-engine; When the lesion location is either in the pancreatic body or the pancreatic tail, the lesion parameters are the lesion location, the distance between the lesion and the pancreatic duct, and the distance between the lesion and the bile duct; when the distance between the lesion and the pancreatic duct or the distance between the lesion and the bile duct is less than or equal to the second threshold, a spleen-preserving pancreatic body and tail resection auxiliary decision is generated based on the second sub-engine.

7. The system according to claim 1, characterized in that, The location of the lesion is determined by the processor executing the following steps: Based on the three-dimensional medical image of the pancreas, the pancreas in the three-dimensional medical image is extracted into a skeleton to obtain the skeleton center point of the pancreas. The skeleton center point is sorted according to a first preset rule to form a skeleton center point set. According to a second preset rule, the skeleton center point set is divided into a first subset, a second subset, a third subset, and a fourth subset. Obtain the centroid of the lesion, calculate the distance between the centroid of the lesion and each skeleton center point in the set of skeleton center points, and determine the skeleton center point closest to the centroid of the lesion as the center point of the lesion location. When the center point of the lesion is located in the first subset, the lesion is located in the head of the pancreas; when the center point of the lesion is located in the second subset, the lesion is located in the neck of the pancreas; when the center point of the lesion is located in the third subset, the lesion is located in the body of the pancreas; when the center point of the lesion is located in the fourth subset, the lesion is located in the body of the pancreas.

8. A method for assisting decision-making regarding pancreatic lesion resection procedures, characterized in that, Acquire three-dimensional medical images of the pancreas, wherein the three-dimensional medical images of the pancreas include lesions; Identify the number and nature of lesions, wherein the number of lesions is one or more, and the nature of the lesions includes at least one of a first lesion type and a second lesion type; A decision tree model is constructed based on the number of lesions, the nature of the lesions, and the surgical procedure function, including: The root node of the decision tree model is constructed based on the number of lesions. When there is only one lesion, the first and second level nodes of the decision tree model are constructed based on the nature of the lesion, and the first and third level nodes of the decision tree model are constructed based on the first formula function, or the second and third level nodes of the decision tree model are constructed based on the second formula function. When there are multiple lesions, the second and second level nodes of the decision tree model are constructed based on the third formula function. The decision tree model is used to generate an auxiliary decision on the surgical procedure for the lesion, including: When there is only one lesion, the first second-level node is executed to determine the nature of the lesion. When the nature of the lesion is the first lesion type, the first third-level node is executed to calculate the output value of the first surgical function. The first surgical procedure function is constructed based on the parameters of the first lesion. The first surgical procedure matching engine includes a first sub-engine, which is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection procedure. The first resection procedure auxiliary decision is generated based on the first sub-engine. The first lesion parameter includes at least one of the lesion location and the distance between the lesion and the blood vessel. The lesion location includes the lesion being located in the pancreatic head, the pancreatic neck, the pancreatic body, or the pancreatic tail. The distance between the lesion and the blood vessel includes at least one of the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein. The first resection procedure auxiliary decision includes: pancreaticoduodenectomy, combined pancreatic head and body resection, pancreatic body and tail resection, and extended pancreatic body and tail resection. When there is only one lesion, the first second-level node is executed to determine the nature of the lesion. When the nature of the lesion is the second lesion type, the second third-level node is executed to calculate the output value of the second surgical function. The second surgical procedure function is constructed based on the second lesion parameters. The first surgical procedure matching engine includes a second sub-engine, which is used to characterize the mapping relationship between the output value of the second surgical procedure function and the resection procedure. The second resection procedure auxiliary decision is generated based on the second sub-engine. The second lesion parameters include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel. The lesion location includes the lesion being located in the pancreatic head, pancreatic neck, pancreatic body, or pancreatic tail. The distance between the lesion and a blood vessel includes at least one of the following: distance between the lesion and the splenic artery, distance between the lesion and the superior mesenteric artery, distance between the lesion and the splenic vein, distance between the lesion and the superior mesenteric vein, and distance between the lesion and the portal vein. The second resection procedure auxiliary decision includes: local resection, pancreatic head resection with preservation of the duodenum, pancreatic head resection with preservation of the duodenum, mid-pancreatic resection with preservation of the duodenum, combined pancreatic head and body resection, and pancreatic body and tail resection with preservation of the spleen. When there are multiple lesions, the second and second-level nodes are executed to calculate the output value of the third surgical procedure function. The output value of the third surgical procedure function is related to the location of the lesion. Based on the output value of the third surgical procedure function, a third resection surgical procedure auxiliary decision is generated for the lesion. The third resection surgical procedure auxiliary decision includes: combined pancreatic head and body resection, pancreatic body and tail resection, and total pancreatectomy.

9. A decision-making aid device for pancreatic lesion resection surgery, characterized in that, The device includes: A medical image acquisition module is used to acquire three-dimensional medical images of the pancreas, wherein the three-dimensional medical images of the pancreas include lesions; The lesion identification module is used to identify the number and nature of lesions, wherein the number of lesions is one or more, and the nature of the lesions includes at least one of a first lesion type and a second lesion type; The decision support module is used to construct a decision tree model based on the number of lesions, the nature of the lesions, and the surgical procedure function, including: The root node of the decision tree model is constructed based on the number of lesions. When there is only one lesion, the first and second level nodes of the decision tree model are constructed based on the nature of the lesion, and the first and third level nodes of the decision tree model are constructed based on the first formula function, or the second and third level nodes of the decision tree model are constructed based on the second formula function. When there are multiple lesions, the second and second level nodes of the decision tree model are constructed based on the third formula function. It is also used to generate surgical intervention decisions corresponding to the lesion based on the decision tree model, including: When there is only one lesion, the first second-level node is executed to determine the nature of the lesion. When the nature of the lesion is the first lesion type, the first third-level node is executed to calculate the output value of the first surgical function. The first surgical procedure function is constructed based on the parameters of the first lesion. The first surgical procedure matching engine includes a first sub-engine, which is used to characterize the mapping relationship between the output value of the first surgical procedure function and the resection procedure. The first resection procedure auxiliary decision is generated based on the first sub-engine. The first lesion parameter includes at least one of the lesion location and the distance between the lesion and the blood vessel. The lesion location includes the lesion being located in the pancreatic head, the pancreatic neck, the pancreatic body, or the pancreatic tail. The distance between the lesion and the blood vessel includes at least one of the distance between the lesion and the splenic artery, the distance between the lesion and the superior mesenteric artery, the distance between the lesion and the splenic vein, the distance between the lesion and the superior mesenteric vein, and the distance between the lesion and the portal vein. The first resection procedure auxiliary decision includes: pancreaticoduodenectomy, combined pancreatic head and body resection, pancreatic body and tail resection, and extended pancreatic body and tail resection. When there is only one lesion, the first second-level node is executed to determine the nature of the lesion. When the nature of the lesion is the second lesion type, the second third-level node is executed to calculate the output value of the second surgical function. The second surgical procedure function is constructed based on the second lesion parameters. The first surgical procedure matching engine includes a second sub-engine, which is used to characterize the mapping relationship between the output value of the second surgical procedure function and the resection procedure. The second resection procedure auxiliary decision is generated based on the second sub-engine. The second lesion parameters include at least one of the following: lesion location, distance between the lesion and the pancreatic duct, distance between the lesion and the bile duct, and distance between the lesion and a blood vessel. The lesion location includes the lesion being located in the pancreatic head, pancreatic neck, pancreatic body, or pancreatic tail. The distance between the lesion and a blood vessel includes at least one of the following: distance between the lesion and the splenic artery, distance between the lesion and the superior mesenteric artery, distance between the lesion and the splenic vein, distance between the lesion and the superior mesenteric vein, and distance between the lesion and the portal vein. The second resection procedure auxiliary decision includes: local resection, pancreatic head resection with preservation of the duodenum, pancreatic head resection with preservation of the duodenum, mid-pancreatic resection with preservation of the duodenum, combined pancreatic head and body resection, and pancreatic body and tail resection with preservation of the spleen. When there are multiple lesions, the second and second-level nodes are executed to calculate the output value of the third surgical procedure function. The output value of the third surgical procedure function is related to the location of the lesion. Based on the output value of the third surgical procedure function, a third resection surgical procedure auxiliary decision is generated for the lesion. The third resection surgical procedure auxiliary decision includes: combined pancreatic head and body resection, pancreatic body and tail resection, and total pancreatectomy.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method of claim 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method of claim 8.

Images