A liver 3D printing model construction method, system and device based on image data

Abstract

This invention provides a method, system, and device for constructing a 3D-printed liver model based on image data, comprising: acquiring CT and MRI image data of an individual patient; performing multimodal image segmentation on the CT and MRI image data to obtain structural segmentation images; fusing the structural segmentation images to obtain a liver and tumor anatomical structure model; performing three-dimensional reconstruction on the liver and tumor anatomical structure model to obtain a target three-dimensional visualization model; and 3D printing the target three-dimensional visualization model. This application integrates CT and MRI imaging modalities to obtain more comprehensive and accurate liver anatomical information. CT provides clear images of the liver parenchyma and blood vessels, while MRI provides more sensitive soft tissue contrast and tumor detection. The combination of the two can compensate for the shortcomings of a single imaging technology, improving the realism and information content of the model.

Description

Technical Field

[0001] This invention relates to the field of medical imaging, specifically to a method, system, and device for constructing a 3D printed liver model based on imaging data. Background Technology

[0002] Transforming medical imaging data into 3D-printed models can not only assist doctors in precise preoperative planning and intraoperative navigation, but also be used in medical education to improve medical students' and residents' understanding of liver anatomy and diseases. Existing research has utilized 3D printing technology to create normal liver models for medical teaching and training. These models can demonstrate the liver's anatomical morphology, Couinaud divisions, vascular distribution, and other three-dimensional structural features. Furthermore, by assembling, disassembling, and performing sectional incisions, learners can observe and understand the liver's internal structure comprehensively and at multiple levels, overcoming to some extent the limitations of traditional cadaver dissection and two-dimensional atlas teaching.

[0003] However, most existing 3D-printed liver models are based on imaging data of healthy individuals and cannot simulate liver tumors. Liver tumors exhibit significant individual differences in size, location, shape, and blood supply, and their relationship with surrounding normal liver tissue and blood vessels also varies. The lack of tumor representation in 3D-printed models makes it difficult to meet the needs of clinical diagnosis and teaching. On the other hand, current clinical diagnosis and surgical planning for liver tumors mainly rely on imaging examinations such as CT and MRI. However, these imaging data are two-dimensional planar images, which cannot intuitively and three-dimensionally display the spatial structural relationship between the liver and tumors, thus failing to meet the needs of clinical diagnosis and teaching. Summary of the Invention

[0004] This application provides a method, system, and device for constructing a 3D printed liver model based on image data, which can intuitively and three-dimensionally display the spatial structural relationship between the liver and tumors to meet the needs of clinical diagnosis and teaching.

[0005] The first aspect of this application provides a method for constructing a 3D printed liver model based on image data;

[0006] A method for constructing a 3D printed liver model based on image data includes:

[0007] Acquire individual patient CT and MRI imaging data;

[0008] Multimodal image segmentation is performed on CT and MRI image data to obtain structural segmentation images;

[0009] By fusing the segmented images, an anatomical model of the liver and tumor is obtained;

[0010] A three-dimensional visualization model of the target is obtained by reconstructing the anatomical structure model of the liver and tumor.

[0011] 3D printing of the target 3D visualization model.

[0012] Preferably, multimodal image segmentation is performed on CT image data and MRI image data to obtain segmented images, including:

[0013] Threshold segmentation is performed on CT image data according to multiple preset threshold ranges to obtain threshold structure segmentation images. Region of interest is extracted from MRI image data using the region growing method to obtain interest structure segmentation images. The structure segmentation images include interest structure segmentation images and threshold structure segmentation images.

[0014] Preferably, before performing multimodal image segmentation on CT and MRI image data to obtain structural segmentation images, the following steps are included:

[0015] Adjust the resolution, pixel size, and coordinate system of the CT and MRI image data to be consistent.

[0016] Preferably, the process of fusing structural segmentation images to obtain a liver and tumor anatomical model includes:

[0017] By inputting the structural segmentation images into a deep learning image fusion model, an anatomical model of the liver and tumor is obtained.

[0018] Preferably, the deep learning image fusion model is a convolutional neural network model or a generative adversarial network model.

[0019] Preferably, a three-dimensional visualization model of the target is obtained by three-dimensional reconstruction of the liver and tumor anatomical structure model, including:

[0020] The deep learning-based image segmentation algorithm segments the organ structure of the liver and tumor anatomical structure model.

[0021] Deep learning-based image feature extraction algorithms extract three-dimensional features of organ structures;

[0022] A three-dimensional visualization model of the target is generated based on the extracted three-dimensional features of the organ and a three-dimensional reconstruction algorithm.

[0023] Preferably, after obtaining the target three-dimensional visualization model by three-dimensional reconstruction of the liver and tumor anatomical structure model, the process includes:

[0024] The target 3D visualization model is smoothed, denoised, and simplified.

[0025] Preferably, after reconstructing the target three-dimensional visualization model from the liver and tumor anatomical structure model, the following are included:

[0026] Convert the target 3D visualization model to STL or OBJ format.

[0027] This application also provides a liver 3D printing model construction system based on image data;

[0028] A liver 3D printing model construction system based on image data includes:

[0029] The image data acquisition module is used to acquire individual patient CT and MRI image data.

[0030] The image segmentation module is used to perform multimodal image segmentation on CT and MRI image data to obtain structural segmentation images.

[0031] The image fusion module is used to fuse structural segmentation images to obtain anatomical models of the liver and tumors;

[0032] The 3D reconstruction module is used to perform 3D reconstruction of the liver and tumor anatomical structure models to obtain the target 3D visualization model.

[0033] The printing module is used for 3D printing of the target 3D visualization model.

[0034] An electronic device is provided in the third aspect of this application.

[0035] An electronic device includes a processor, a memory, and a transceiver. The memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform a method for constructing a liver 3D printing model based on image data.

[0036] In summary, this application includes the following beneficial technical effects:

[0037] 1. This application can provide a realistic, intuitive, and tangible physical model of the liver, which helps to explain the condition to patients and their families before and after surgery.

[0038] 2. This application uses partition marking and simulates surgical procedures on the model, which can improve the accuracy of preoperative planning and reduce surgical risks.

[0039] 3. This application can predict liver function and guide clinical decision-making by calculating the resected and residual liver volume.

[0040] 4. The liver-like tissue structure produced by bio-3D printing in this application can be used for in vitro experimental studies such as drug metabolism.

[0041] 5. Vascular network reconstruction technology can display the three-dimensional blood supply of the liver, and more accurately determine the relationship between the tumor and the surrounding blood vessels.

[0042] 6. This application integrates CT and MRI imaging modalities to obtain more comprehensive and accurate liver anatomical information. CT provides clear images of the liver parenchyma and blood vessels, while MRI provides more sensitive soft tissue contrast and tumor detection. The combination of the two can compensate for the shortcomings of a single imaging technology and improve the realism and information content of the model.

[0043] 7. This application establishes a liver tumor model and uses a laparoscopic system to train laparoscopic ultrasound operation, enabling hepatobiliary surgeons to master laparoscopic ultrasound technology and perform puncture tumor simulation ablation under laparoscopic ultrasound guidance, thereby shortening the learning curve. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating an embodiment of a liver 3D printing model construction method based on image data according to this application;

[0045] Figure 2 This is CT image data from an embodiment of a liver 3D printing model construction method based on image data according to this application;

[0046] Figure 3 This is MRI image data from an embodiment of a liver 3D printing model construction method based on image data according to this application;

[0047] Figure 4 This is a schematic diagram of the target three-dimensional visualization model from a first-view perspective of an embodiment of a liver 3D printing model construction method based on image data according to this application;

[0048] Figure 5 This is a schematic diagram of the target three-dimensional visualization model from a second perspective, representing an embodiment of a liver 3D printing model construction method based on image data according to this application.

[0049] Figure 6 This is a schematic diagram of the target 3D visualization model from a third-person perspective, representing an embodiment of a liver 3D printing model construction method based on image data according to this application.

[0050] Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0051] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0052] In the description of the embodiments in this application, words such as "illustrative," "for example," or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "illustrative," "for example," or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of words such as "illustrative," "for example," or "for example" is intended to present the relevant concepts in a specific manner.

[0053] In the description of the embodiments of this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, B existing alone, or A and B existing simultaneously. Furthermore, unless otherwise stated, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and their variations all mean "including but not limited to," unless otherwise specifically emphasized.

[0054] Scenario: To overcome the limitations of traditional cadaver dissection and two-dimensional atlas teaching, researchers used 3D printing technology to create normal liver models for medical teaching and training. Most existing 3D-printed liver models are based on imaging data of normal individuals and cannot simulate liver tumors. Liver tumors exhibit significant individual differences in size, location, shape, and blood supply, and their relationship with surrounding normal liver tissue and blood vessels also varies. The lack of realistic tumor representation in 3D-printed models makes it difficult to fully meet the needs of clinical diagnosis and teaching.

[0055] To address this issue, this application uses the patient's CT and MRI image data as a basis, employing image segmentation and 3D reconstruction techniques to accurately extract the 3D structural information of the liver and tumor. This information is then combined with 3D printing technology to create a highly realistic, multi-material, and detachable liver tumor model. This method not only realistically recreates the complex structure of the liver and its tumor but also provides precise preoperative planning and guidance.

[0056] Explanation of proper nouns:

[0057] Region growing is an image segmentation technique based on the principle of image homogeneity, meaning that similar pixels in an image tend to belong to the same region. Region growing starts with a set of seed points and gradually adds adjacent pixels to the regions represented by these seed points until no more pixels can be added. The basic steps of region growing typically include: selecting seed points: choosing a set of pixels in the image as seeds for the initial regions; defining similarity criteria: determining the criteria for whether a pixel should be added to an existing region, which usually involves attributes such as pixel grayscale value, texture, and color; the growing process: starting from a seed point, examining its surrounding pixels. If these pixels meet the similarity criteria, they are added to the corresponding region, becoming new seed points; and repeating the growing process: repeating the above process until all possible pixels have been added to the corresponding regions, or a certain termination condition is reached.

[0058] Mimics: A modular software that can be used in various combinations to meet different user needs. Mimics offers a variety of functions, including but not limited to: Image Import and Segmentation: Supports importing various image formats and provides multiple segmentation tools such as grayscale thresholding and region growing to help users quickly highlight regions of interest. Image Visualization: Provides axial, coronal, and sagittal views of the raw data, as well as a 3D view reconstructed from the region of interest, supporting translation, scaling, and rotation of the 3D view. Image Registration and Measurement: Features image registration, point registration, and STL registration functions, as well as point-to-point measurement, contour line measurement, and grayscale value measurement.

[0059] This application provides a method for constructing a liver 3D printing model based on image data;

[0060] Reference Figure 1 A method for constructing a 3D printed liver model based on image data, specifically including the following steps:

[0061] S1: Acquire individual patient CT and MRI image data;

[0062] Specifically, this includes CT image data acquisition and MRI image data acquisition:

[0063] S11: CT image data acquisition;

[0064] Reference Figure 1 and Figure 2A 64-slice 128-layer spiral CT scanner (such as GE LightSpeed ​​VCT) was used with a scanning voltage of 120kV, tube current of 300mAs, acquisition frequency of 0.5s / rot, and pitch of 1.375:1 to acquire thin-slice CT images of the patient's upper abdomen in the form of 1.0-1.5mm slices. The scanning range was from the upper edge to the lower edge of the liver. During the scan, four phases of images were acquired simultaneously: plain phase, arterial phase (25-30s after enhancement), portal venous phase (50-60s), and equilibrium phase (100-120s). The obtained CT image data were stored in DICOM 3.0 format.

[0065] S12: MRI image data acquisition;

[0066] Reference Figure 1 and Figure 3 Using a 1.5T or 3.0T MRI scanner (such as the Siemens Skyra), a voxel-isotropic three-dimensional rapid acquisition gradient echo sequence (3D VIBE) was employed, with a TR of 3.9 ms, TE of 1.4 ms, a flip angle of 9 degrees, and a slice thickness of 2.0-3.0 mm to acquire T1WI dynamically enhanced MRI images. Images were acquired during the arterial, portal venous, and equilibrium phases after gadopentetate dimeglumine injection. Simultaneously, conventional T2WI sequences were performed to acquire anatomical structure images. The resulting images were also stored in DICOM format.

[0067] Specifically, personnel can adjust and optimize the data acquisition process according to the actual situation. For example, the scanning parameters can be optimized, such as adjusting the tube current to 200-400 mAs, the acquisition frequency to 0.4-0.6 s / rot, and the pitch to 1.2-1.5:1, in order to balance image quality and radiation dose. The scanning slice thickness can be appropriately adjusted to 0.5-2.0 mm according to the type and size of the liver tumor, in order to balance resolution and data volume. The time points of the four-phase enhanced scanning can be fine-tuned according to individual differences in cycle time to obtain the best vascular imaging effect.

[0068] S2: Adjust the resolution, pixel size, and coordinate system of CT and MRI image data to be consistent;

[0069] Specifically, CT and MRI image data are imported into medical image post-processing software (such as Materialise Mimics). The CT image data is used as a fixed image and the MRI image data is used as a floating image. Preprocessing is performed by manually selecting feature points for coarse registration and resampling. Then, the mutual information algorithm is used for fine registration, so that the two sets of images have the same resolution and pixel size, and the two sets of images are matched to the same coordinate system.

[0070] S3: Perform multimodal image segmentation on CT and MRI image data to obtain structural segmentation images;

[0071] Specifically, steps S31-S32 are included, wherein the structure segmentation image includes an interest structure segmentation image and a threshold structure segmentation image;

[0072] S31: Threshold segmentation is performed on CT image data according to multiple preset threshold ranges to obtain a threshold structure segmentation image;

[0073] Specifically, CT image data is segmented using thresholding to obtain threshold structure segmentation images. Multiple threshold ranges are included: liver parenchyma threshold: 50-150 HU, tumor threshold: 150-300 HU, artery threshold: 250-450 HU, and vein threshold: 100-250 HU. These threshold segmentation parameters can be fine-tuned based on actual image quality and contrast; for example, the threshold ranges for liver parenchyma and tumors can be adjusted to 40-180 HU and 120-350 HU, respectively.

[0074] S32: Extract the region of interest from the MRI image data using the region growing method to obtain the segmentation image of the structure of interest.

[0075] Specifically, the MRI image data processing uses the region growing method, starting with multiple seed points including tumors and blood vessels, to extract regions of interest (ROIs) as segmentation images of structures of interest.

[0076] In other embodiments, vessel segmentation can be combined with region growing and thresholding methods to improve segmentation accuracy and continuity, and manual editing corrections can be performed when necessary.

[0077] It should be noted that after segmenting the image based on the structure of interest and the image based on the threshold structure, this example performs surface smoothing and mesh optimization on regions such as the liver, tumor, and blood vessels in the image to improve the printing quality of the 3D model.

[0078] S4: Fusion of structural segmentation images to obtain anatomical models of the liver and tumors;

[0079] Specifically, the structural segmentation images are input into a deep learning image fusion model to obtain a liver and tumor anatomical structure model. The deep learning image fusion model is either a convolutional neural network (CNN) or a generative adversarial network (GAN) model. This model has been trained to automatically learn the correlation features between the interest-based structural segmentation images and the threshold-based structural segmentation images. The training of this deep learning image fusion model follows existing neural network training processes; it can be understood as existing technology, which will not be elaborated upon here. It should be noted that the training process of the deep learning image fusion model can involve inputting the interest-based structural segmentation images and the threshold-based structural segmentation images into the deep learning image fusion model to generate a high-quality fused image, i.e., the liver and tumor anatomical structure model.

[0080] It is worth noting that steps S3-S4 also include a vascular network reconstruction step, specifically:

[0081] Based on arterial and venous phase data from CT imaging, Mimics reconstructs vascular networks in response to human intervention. Hepatic arteries are extracted using a threshold of 250-450 HU, and hepatic veins are extracted using 100-250 HU. Manual editing and correction are performed, and a 1-pixel brush tool is used to ensure the continuity of the vessels. The reconstructed hepatic artery, portal vein, and hepatic vein networks are then colored to show their three-dimensional course and their positional relationship with the liver and tumor. Finally, the network is fused with a model of the liver and tumor anatomy to generate a transparent, overall liver vascular network model.

[0082] The specific steps are as follows:

[0083] 1. Perform vessel segmentation on CT images of the arteriovenous phase, and select the threshold according to the method described above;

[0084] 2. Morphological processing of the segmentation results was performed, and 1×1×1 structural elements were used to expand the blood vessels and fill the ruptured lumen.

[0085] 3. The portal vein and hepatic vein are fused together, and the surface is smoothed to generate a complete liver vascular tree;

[0086] 4. Register with the liver parenchyma and tumor model, i.e., the liver and tumor anatomical structure model, to obtain a complete liver model containing a fine vascular network.

[0087] The vascular reconstruction threshold can be adjusted based on the quality of vascular visualization, generally >300 HU for arteries and >150 HU for veins. Vascular continuity processing can be combined with morphological algorithms and artificial intelligence technologies to improve automation and efficiency. The positional relationship between the vascular network and the liver and tumors can be visualized using color coding, transparent rendering, and other methods.

[0088] S5: The target three-dimensional visualization model is obtained by reconstructing the anatomical structure model of the liver and tumor.

[0089] Specifically, refer to Figure 1 and Figure 4In Mimics software, three-dimensional reconstructions are performed on the liver parenchyma, tumor, hepatic artery, hepatic vein, and portal vein structures in the liver and tumor anatomical model. The organ structures in the liver and tumor anatomical model can be segmented using deep learning-based image segmentation algorithms. These deep learning-based image segmentation algorithms, such as U-Net and V-Net, can automatically segment the target organs. The segmentation network trained by the deep learning image segmentation algorithm learns the feature representations of organ boundaries and internal structures through training on a large amount of labeled data, enabling it to quickly and accurately extract the organ regions of interest. The segmentation network trained by the deep learning image segmentation algorithm can be an existing pre-trained neural network model.

[0090] In Mimics software, the region growing method is used to group the masks of various organ structures, converting the masks into 3D objects. The generated target 3D visualization model undergoes post-processing and optimization, such as smoothing, denoising, and simplification, to better suit the requirements of 3D printing. Editing tools are used to smooth the 3D model, eliminating burrs and sharp corners. (Refer to...) Figure 5 and Figure 6 The transparency of the 3D model is adjusted to reveal its internal structure. The liver parenchyma, tumor, blood vessels, etc., are spatially registered according to a preset anatomical order to obtain a complete 3D visualization model of the liver.

[0091] Methods for converting a mask into a 3D object include: volume rendering, maximum density projection, stereo density projection, minimum density projection, or surface maximum density projection.

[0092] In this embodiment, the segmented liver, tumor, and blood vessel masks are reconstructed in 3D using the Marching Cubes algorithm to generate a smooth, defect-free STL format model file; in other embodiments, they can also be converted to OBJ format model files.

[0093] In another embodiment, the three-dimensional features of organ structures can be extracted based on deep learning image feature extraction algorithms. The deep learning image feature extraction algorithm can adopt a three-dimensional CNN algorithm that automatically learns the high-level semantic features of organs, or it can adopt traditional image feature extraction algorithms, such as gray-level co-occurrence matrix (GLCM), wavelet transform, etc., to replace the deep learning image feature extraction algorithm for extracting the three-dimensional features of organ structures.

[0094] Then, based on the extracted organ 3D features and 3D reconstruction algorithms, a target 3D visualization model is generated. Specific 3D reconstruction algorithms include Marching Cubes or Poisson Surface Reconstruction. The extracted organ 3D features are processed by the 3D reconstruction algorithm to generate the target 3D visualization model.

[0095] In other embodiments, deep learning-based 3D generative models, such as 3D GANs, can be used to directly generate a 3D visualization model of the target from fused images and feature representations.

[0096] After the target 3D visualization model is generated, it undergoes smoothing, denoising, and simplification processes to make it more suitable for 3D printing requirements. The processing techniques used can be traditional mesh processing algorithms or deep learning-based 3D model optimization methods.

[0097] To facilitate teaching and simulation demonstrations, the 3D-printed liver model was divided into zones I-VIII, referencing the Couinaud segmentation system, and each zone was labeled with a different color, as follows:

[0098] Zone I: Tail-shaped leaves, yellow

[0099] Zone II: Left lateral lobe posterior superior region, green

[0100] Zone III: Left lateral lobe anterior superior region, orange

[0101] Area IVa: Left inner lobe supraoral region, purple

[0102] Area IVb: Inferior region of the left inner lobe, brown

[0103] Section V: Lower right front section, blue

[0104] Zone VI: Lower right rear area, indigo blue

[0105] Section VII: Upper right rear section, red

[0106] Section VIII: Upper right front section, pink

[0107] Furthermore, based on Couinaud's eight segments, it is further subdivided into subsegments, dividing segment VIII into two subsegments, VIIIa and VIIIb, which respectively represent the superior and inferior drainage areas of the segment VIII vein, thus reflecting the intrahepatic vascular distribution in a more precise way.

[0108] The tumor was located using CT imaging data, its relationship with surrounding blood vessels was marked, and the resection path and extent were planned. Mimics' CAD measurement tool was used to measure the volume of the liver to be resected and the volume of the remaining liver, and to assess postoperative liver function reserve.

[0109] Tumor localization and vascular relationship marking should refer to imaging diagnosis, and transparent processing or section analysis may be performed when necessary. Surgical planning should comprehensively consider factors such as tumor size, location, affected liver segment, and residual liver volume, and 3D printing can be used to simulate surgical procedures.

[0110] S5: 3D printing of the target 3D visualization model.

[0111] Specifically, 3D printing preprocessing is required before 3D printing;

[0112] 3D printing preprocessing:

[0113] The reconstructed 3D liver model is exported in STL format and imported into 3D printing slicing software (such as Cura or Slic3r) for preprocessing before printing. Printing parameters are adjusted according to the model's size and shape.

[0114] Print size: scaled proportionally to the original model to a 1:1 scale;

[0115] Printing materials: PLA, ABS or photosensitive resin can be selected, and different printing temperatures and speeds can be set according to the material properties;

[0116] Layer thickness: 0.1-0.2mm. The thinner the layer, the higher the printing accuracy, but the longer the printing time.

[0117] Infill rate: 10%-30%. The higher the infill rate, the stronger the model, but the more printing time and material are required.

[0118] Support structure: Add support structures automatically or manually according to the overhang and concavity of the model to prevent printing deformation;

[0119] Printing direction: Adjust the printing direction according to the structural characteristics of the model to optimize the printing effect and support arrangement.

[0120] 3D printing is performed after pre-processing. Specifically:

[0121] Transfer the processed data file from the slicing software to the 3D printer (such as a Stratasys Fortus 450mc). Select the appropriate printing material and load it into the printer, then calibrate the level of the printing platform. Start the printing program; the printer will print layer by layer according to the preset path. After each layer is printed, the platform will descend one layer height until the entire model is printed. Adjust the printing parameters as needed during the printing process and monitor the printer in real time. Spray release agent appropriately during printing, and after printing, soak the model in release fluid to dissolve the support material.

[0122] Print size: scaled proportionally to the original model to a 1:1 scale;

[0123] Printing materials: PLA, ABS, or photosensitive resin;

[0124] Layer thickness: 0.1-0.2mm;

[0125] Fill rate: 10%-30%;

[0126] Printing speed: 50-80mm / s;

[0127] Printing temperature: PLA 200-220℃, ABS 230-250℃;

[0128] Heated bed temperature: PLA 50℃, ABS 80-100℃;

[0129] Transfer the sliced ​​data file to a 3D printer (such as a Stratasys Fortus 450mc) and begin printing.

[0130] The printing material can also be a flexible material such as TPU to simulate the feel of a liver.

[0131] (4) Post-printing processing

[0132] After printing, remove the printed part from the printing platform and carefully remove the support material. Use scissors, scalpels, or other tools to cut away the support, then sand the model surface until smooth. Color the models of different tissue structures, such as brownish-yellow for liver parenchyma, red for hepatic arteries, blue for hepatic veins and portal veins, and green for tumors, to enhance structural distinctiveness. Assemble the liver parenchyma and blood vessels to form a complete liver model. Finally, spray with a clear protective varnish to enhance the model's strength and durability.

[0133] Various surface treatment techniques, such as grinding, polishing, painting, and electroplating, can be used to improve the model's aesthetics and durability. For intricate vascular structures, staining and water transfer printing can be used for coloring. The liver model can be assembled using detachable connection methods such as snap-fit ​​and magnetic attachment, facilitating repeated use and replacement.

[0134] It should also be noted that:

[0135] This application also optimizes the preparation and configuration of 3D printing materials, specifically as follows:

[0136] Preparation of bio-3D printing materials:

[0137] Gelatin, sodium alginate, and fibrinogen were mixed in a mass ratio of 8:1:1 and prepared into a hydrogel ink with a final concentration of 10% using phosphate-buffered saline (PBS). The HepG2 human liver cell line was isolated and cultured, added to the hydrogel at a cell density of 10⁷ cells / ml, and stirred at low temperature until homogeneous, thus obtaining the cell-hydrogel hybrid ink (Bio-ink) for bio-3D printing.

[0138] Load the Bio-ink into the print head of a bio-3D printer (such as the Organovo Novogen MMX) and set the printing parameters:

[0139] Nozzle diameter: 200-500μm

[0140] Air pressure: 0.1-0.3 MPa

[0141] Printing speed: 5-20mm / s

[0142] Layer thickness: 100-300μm

[0143] A layer of CaCl2 solution was spread on the collection plate as a cross-linking agent, and the cell-hydrogel scaffold was printed layer by layer. After printing, the scaffold was cross-linked and solidified with CaCl2 and glutaminase solution, washed with PBS, and then incubated in a 37°C, 5% CO2 incubator. After 2 weeks, a liver-like tissue structure was formed.

[0144] The specific steps for bio-3D printing are as follows:

[0145] ① Cell preparation: HepG2 hepatocytes were isolated and cultured, and the cell density was adjusted to 107 cells / ml;

[0146] ② Scaffold design: Inspired by the hexagonal structure of liver lobules, a honeycomb scaffold with a central lumen is designed with a porosity >90% and a pore size of 100-500μm;

[0147] ③ Material preparation: Gelatin, sodium alginate and fibrinogen were mixed in a ratio of 8:1:1, dissolved in PBS to prepare a 10% hydrogel, and HepG2 cell suspension was added to prepare Bioink.

[0148] ④ 3D printing: Using a pressure extrusion bio-3D printer with a nozzle inner diameter of 400μm, an air pressure of 0.1MPa, a printing speed of 10mm / s, and a layer thickness of 200μm, a cell-scaffold composite structure was printed.

[0149] ⑤ Post-processing: Cross-linked with CaCl2 and glutaminase, washed with PBS, and cultured at 37℃ and 5% CO2 for 2 weeks to form liver-like tissue;

[0150] ⑥ Characterization and evaluation: Hematoxylin-eosin staining was used to observe histological morphology, Coomassie brilliant blue and PAS staining were used to assess glycogen and lipid content, and albumin and urea synthesis tests were used to detect hepatocyte function.

[0151] The hydrogel material ratio can be adjusted as needed, such as increasing the fibrinogen ratio to enhance scaffold strength. Cell density can be optimized based on tissue type and printing parameters, typically ranging from 10⁵ to 10⁸ cells / ml. Cell culture conditions, such as culture medium composition, pH, and temperature, should be optimized according to cell type.

[0152] 3D printing material configuration:

[0153] (1) Hepatocyte preparation: Hepatocytes are cultured in the logarithmic growth phase and then digested and collected.

[0154] (2) Material preparation: Mix the sterilized 4% sodium alginate, 20% gelatin and 3% fibrinogen in a ratio of 1:1:1 and dissolve for later use;

[0155] (3) Preparation of cell-loaded hydrogel scaffold: The hepatocyte suspension and glutamine transaminase were mixed at a ratio of 6:1. The cell mixture with added TGaes was mixed evenly with the rewarmed sodium alginate / gelatin / fibrinogen hydrogel system solution and loaded into a syringe for pre-cooling.

[0156] (4) 3D bioprinting: The pre-cooled sodium alginate / gelatin / fibrinogen mixture containing cell suspension is installed into the 3D bioprinter and printed according to the pre-programmed printing route.

[0157] (5) Post-printing processing: After printing, cross-linking is performed using 3% CaCl2, followed by the addition of thrombin to cross-link fibrinogen, and then cultured in complete culture medium.

[0158] (6) Sodium alginate saline gel removal: EDTA + sodium citrate were used to remove the sodium alginate shell to obtain hepatocytes grown in a 3D in vitro model.

[0159] In bio-3D printing, hydrogel materials can be selected from different natural or synthetic polymers depending on the cell type and printing requirements. The mixing ratio of cell suspension and hydrogel can be optimized based on material properties and cell density, generally with a cell volume fraction of 10%-50%. Printing parameters such as nozzle diameter, air pressure, and printing speed can be adjusted according to the rheological properties of the material and cell activity. Crosslinking methods can be selected based on material properties, including ionic crosslinking, enzymatic crosslinking, and photocrosslinking, with optimized crosslinking time and conditions. Shell removal methods can include dissolution and enzymatic hydrolysis to avoid affecting cell activity.

[0160] Furthermore, based on the CT and MRI imaging characteristics of primary liver cancer, metastatic liver cancer, and hepatic hemangioma, this application designs three 3D printing schemes to simulate their morphology and internal structure, respectively, in order to achieve realistic 3D printed models.

[0161] (1) 3D printing solution for primary liver cancer:

[0162] The irregular shape and uneven surface of the l reflect the invasive growth of the tumor.

[0163] The internal material has uneven density, and low-density areas are printed locally to simulate tumor necrosis, cystic degeneration, and fatty degeneration.

[0164] High-density materials are used locally in a patchy distribution to simulate intratumoral hemorrhage.

[0165] A dense, thin layer is printed around the tumor to simulate a pseudocapsule.

[0166] If necessary, dendritic structures can be printed within the lesion to simulate blood vessels inside the tumor.

[0167] (2) 3D printing solutions for metastatic liver cancer:

[0168] Using regular shapes such as spheres or ellipses, multiple lesions are printed in a dispersed manner to simulate multiple metastatic lesions.

[0169] The interior uses a uniform low-density material to reflect the homogeneity of metastatic tumors.

[0170] The central area is printed with an irregular low-density zone, surrounded by a high-density ring, forming a "bull's eye sign".

[0171] A blurred, high-density ring is printed around the lesion to simulate a "halo sign".

[0172] In some lesions, a very low-density area can be printed in the center, surrounded by a high-density ring, presenting a "ring target sign".

[0173] (3) 3D printing solution for hepatic hemangioma:

[0174] l has a lobed, oval, or round appearance with clear boundaries.

[0175] The interior uses a homogeneous, low-density material to reflect the spongy structure of hemangiomas.

[0176] In the center of a larger diameter lesion, print star-shaped or slit-shaped ultra-low density areas to simulate thrombosis, bleeding, or scarring.

[0177] Enhanced scanning protocol: First, print a high-density edge layer, then gradually print towards the center until the lesion is filled.

[0178] If necessary, the entire lesion can be printed as an extremely high density in the severe T2WI protocol to simulate the "light-up sign".

[0179] It should also be noted that the CT and MRI images of the aforementioned individual patients, as well as the target 3D visualization model, are all derived from actual patient cases, the basic information of which is as follows:

[0180] The patient is a 61-year-old female. Seven years ago, an abdominal ultrasound during a routine check-up at a local hospital suggested a possible hepatic hemangioma, approximately 2-3 cm in size. The patient had no history of abdominal distension or pain and did not seek medical attention. One week ago, she underwent another abdominal MRI at the same local hospital, which revealed a hepatic hemangioma, approximately 5-7 cm in size. She reported no significant pain, nausea, belching, acid reflux, or choking sensation when eating. She also reported no chills, rigors, or fever during the course of her illness. For further diagnosis and treatment, the outpatient department admitted her to the hospital with a preliminary diagnosis of "hepatic space-occupying lesion."

[0181] The implementation process of the liver 3D printing model construction method based on image data provided in this application includes: after acquiring the CT image data and MRI image data of an individual patient, performing multimodal image segmentation on the CT image data and MRI image data to obtain structural segmentation images; the structural segmentation images include segmentation images of liver tissue structure and tumor structure; the segmentation of the tumor structure is achieved through the following steps: extracting high-density regions as tumor candidate regions in CT images through threshold segmentation; highlighting the differences between the tumor region and normal liver tissue in T1-weighted and T2-weighted MRI images using dynamic contrast enhancement and diffusion-weighted imaging techniques; and further processing the tumor candidate regions extracted from CT and MRI images. Spatial registration is performed, and precise tumor boundaries are extracted through morphological operations and connectivity analysis. Segmented images are fused to obtain anatomical models of the liver and tumor. A deep learning algorithm is used to fuse segmented images of liver tissue and tumor structures, generating complete anatomical models of the liver and tumor. This deep learning algorithm is based on the U-Net architecture, fusing features from CT and MRI modalities in the encoder and reconstructing continuous and smooth 3D structures of the liver and tumor in the decoder. Based on the fused anatomical models, geometric features of the liver and tumor surfaces, as well as topological features of internal liver vessels, bile ducts, and other ductal structures, are extracted. An improved marching algorithm is employed. The Cubes algorithm and Dijkstra's shortest path algorithm are used to reconstruct the surface mesh of the liver and tumor, and extract the skeleton of the duct structure, respectively. The reconstructed liver and tumor surface models and duct structure models are assembled to obtain high-fidelity 3D visualization models of the liver and tumor. According to the clinical needs of doctors, the size, color, transparency, etc. of the visualization model are personalized. The personalized 3D visualization models of the liver and tumor are converted into 3D printing model data. Appropriate printing materials (such as photosensitive resin, silicone, etc.) and printing process parameters are selected to realize the rapid manufacturing of the physical model of the liver and tumor on a 3D printer. After the printed model undergoes necessary post-processing steps such as cleaning and curing, the final 3D printed models of the liver and liver tumor are obtained.

[0182] This application also provides a method for constructing a liver 3D printing model based on image data;

[0183] A liver 3D printing model construction system based on image data includes:

[0184] The image data acquisition module is used to acquire individual patient CT and MRI image data.

[0185] The image segmentation module is used to perform multimodal image segmentation on CT and MRI image data to obtain structural segmentation images.

[0186] The image fusion module is used to fuse structural segmentation images to obtain anatomical models of the liver and tumors;

[0187] The 3D reconstruction module is used to perform 3D reconstruction of the liver and tumor anatomical structure models to obtain the target 3D visualization model.

[0188] The printing module is used for 3D printing of the target 3D visualization model.

[0189] Reference Figure 7 The present application provides a schematic diagram of the structure of an electronic device. The electronic device may include: at least one processor, at least one network interface, a user interface, a memory, and at least one communication bus.

[0190] The communication bus is used to enable communication between these components.

[0191] The user interface may include a display screen and a camera. Optional user interfaces may also include standard wired interfaces and wireless interfaces.

[0192] The network interface may include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0193] The processor may include one or more processing cores. It connects to various parts of the server via various interfaces and lines, executing instructions, programs, code sets, or instruction sets stored in memory, and accessing data stored in memory to perform various server functions and process data. Optionally, the processor may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor may integrate one or more of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip without being integrated into the processor.

[0194] The memory may include random access memory (RAM) or read-only memory. Optionally, the memory may include a non-transitory computer-readable storage medium. The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor. The memory, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for constructing a liver 3D printing model based on image data.

[0195] It should be noted that the above embodiments of the apparatus are only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0196] In electronic devices, the user interface is mainly used to provide an interface for users to input data and obtain user input data; while the processor can be used to call an application stored in memory that describes a method for constructing a liver 3D printing model based on image data. When executed by one or more processors, the electronic device performs one or more of the methods described in the above embodiments.

[0197] In this specification, “unit” and “module” refer to software and / or hardware that can independently or in conjunction with other components perform a specific function. Hardware may include, for example, a Field-Programmable Gate Array (FPGA) or an Integrated Circuit (IC).

[0198] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0199] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0200] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0201] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0202] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0203] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0204] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0205] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will be readily apparent to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for constructing a 3D printed liver model based on image data, characterized in that, include: Acquire individual patient CT and MRI imaging data; Multimodal image segmentation was performed on CT and MRI image data to obtain structural segmentation images; Vascular network reconstruction based on arterial and venous phase data from CT imaging; By fusing the segmented images, an anatomical model of the liver and tumor is obtained; The reconstructed vascular network was registered with the liver and tumor anatomical model to obtain a whole liver model containing a fine vascular network. The overall liver model is reconstructed in three dimensions to obtain a target three-dimensional visualization model; 3D printing of the target 3D visualization model; The reconstruction of the vascular network includes the following steps: Blood vessel segmentation was performed on the arteriovenous phase CT images. The hepatic artery was extracted with a threshold of 250-450 HU, and the hepatic vein was extracted with a threshold of 100-250 HU. Morphological processing was performed on the segmentation results, and 1×1×1 structural elements were used to expand the blood vessels and fill the ruptured lumen. The portal vein and hepatic vein are fused together, and the surface is smoothed to generate a complete liver vascular tree. Referring to the Couinaud segmentation system, the 3D-printed liver model was divided into regions I-VIII, with each region labeled with a different color. Based on the Couinaud eight-segment system, it was further subdivided into subsegments, with region VIII divided into two subsegments, VIIIa and VIIIb, representing the superior and inferior drainage areas of the segment VIII vein, respectively.

2. The method for constructing a liver 3D printing model based on image data according to claim 1, characterized in that, Multimodal image segmentation was performed on CT and MRI image data to obtain segmented images, including: Threshold segmentation is performed on CT image data according to multiple preset threshold ranges to obtain threshold structure segmentation images. Region of interest is extracted from MRI image data using the region growing method to obtain interest structure segmentation images. The structure segmentation images include interest structure segmentation images and threshold structure segmentation images.

3. The method for constructing a liver 3D printing model based on image data according to claim 2, characterized in that, The structural segmentation image includes a segmentation image of the tumor structure, and the segmentation of the tumor structure segmentation image includes: High-density regions in CT images are extracted as tumor candidate regions through threshold segmentation. Dynamic contrast enhancement and diffusion-weighted imaging techniques were used to highlight the differences between tumor regions and normal liver tissue in T1-weighted and T2-weighted MRI images; Tumor candidate regions extracted from CT and MRI images are spatially registered, and the precise boundaries of the tumor are extracted through morphological operations and connectivity analysis to obtain a segmented image of the tumor structure.

4. The method for constructing a liver 3D printing model based on image data according to claim 3, characterized in that, The structural segmentation images include segmented images of liver tissue structures. The process of fusing these segmented images to obtain a liver and tumor anatomical model includes: A deep learning algorithm is used to fuse segmented images of liver tissue structure and tumor structure to generate a complete anatomical model of the liver and tumor. The deep learning algorithm is based on the U-Net architecture.

5. The method for constructing a liver 3D printing model based on image data according to claim 1, characterized in that, 3D printing... include: By selecting appropriate printing materials and printing process parameters, the target three-dimensional visualization model corresponding to the liver and tumor can be rapidly manufactured on a 3D printer. After the printed model is cleaned and cured, the final 3D printed model of the liver and liver tumor is obtained.

6. The method for constructing a liver 3D printing model based on image data according to claim 1, characterized in that, A three-dimensional visualization model of the target liver and tumor anatomical structure was obtained by reconstructing the anatomical structure model, including: The deep learning-based image segmentation algorithm segments the organ structure of the liver and tumor anatomical structure model. Deep learning-based image feature extraction algorithms extract three-dimensional features of organ structures; A three-dimensional visualization model of the target is generated based on the extracted three-dimensional features of the organ and a three-dimensional reconstruction algorithm.

7. The method for constructing a liver 3D printing model based on image data according to claim 1, characterized in that, A three-dimensional visualization model of the target liver and tumor anatomical structure was obtained by reconstructing the anatomical structure model, including: The encoder section fuses features from both CT and MRI modalities, while the decoder section reconstructs a continuous and smooth three-dimensional structure of the liver and tumor. Based on the fused liver and tumor anatomical models, the geometric features of the liver surface and tumor surface, as well as the topological features of the internal liver vessels, bile ducts, and other ductal structures, are extracted. The reconstructed liver, tumor surface model, and duct structure model are assembled to obtain a high-fidelity 3D visualization model of the liver and tumor; the size, color, transparency, and other settings of the visualization model are personalized according to the doctor's clinical needs. The personalized 3D visualization models of the liver and tumors are converted into 3D printable model data.

8. The method for constructing a liver 3D printing model based on image data according to claim 7, characterized in that, The extraction of geometric features of the liver and tumor surfaces, as well as topological features of ductal structures such as blood vessels and bile ducts inside the liver, includes: using the Marching Cubes algorithm and the Dijkstra shortest path algorithm to reconstruct the meshes on the liver and tumor surfaces and extract the skeleton of the ductal structures, respectively.

9. A system for constructing a 3D printed liver model based on imaging data obtained by the method described in any one of claims 1-8, characterized in that, include: The image data acquisition module is used to acquire individual patient CT and MRI image data. The image segmentation module is used to perform multimodal image segmentation on CT and MRI image data to obtain structural segmentation images. The image fusion module is used to fuse structural segmentation images to obtain anatomical models of the liver and tumors; The 3D reconstruction module is used to perform 3D reconstruction of the liver and tumor anatomical structure models to obtain the target 3D visualization model. The printing module is used for 3D printing of the target 3D visualization model.

10. An electronic device, characterized in that, The device includes a processor, a memory, and a transceiver, wherein the memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-8.

Images