A monitoring method, system, storage medium and electronic device during thyroid ablation surgery

Abstract

The present invention provides a method, system, storage medium and electronic device for monitoring during thyroid ablation. The system includes a data acquisition unit; an image segmentation unit for obtaining thyroid organ segmentation areas and lesion areas, and generating ablation safety areas based on spatial information of the lesion areas; a three-dimensional reconstruction unit for obtaining and reconstructing organ volume data; and a dynamic monitoring unit for real-time registration of preoperative image volume data with ultrasound image sequences obtained during surgery to observe the location of the lesion, obtain thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of ablation. The present invention eliminates the 3D reconstruction process of data collected during surgery through one registration plus multiple fast secondary precision registrations, facilitates monitoring of thyroid deformation under ablation, and solves the problem of unclear positioning of ultrasound lesions during surgery with preoperative and intraoperative registration, thereby obtaining better ablation results.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and more specifically, to a monitoring method, system, storage medium and electronic device during thyroid ablation surgery. Background Art

[0002] Thyroid ablation is a minimally invasive procedure used to treat thyroid nodules, hyperthyroidism, and thyroid cancer. The main techniques include radiofrequency ablation (RFA), microwave ablation (MWA), laser ablation (LA), and high-intensity focused ultrasound (HIFU).

[0003] However, due to reasons such as thyroid deformation during ablation, glandular displacement caused by water isolation technology, and vaporization during thermal ablation, ultrasound imaging is difficult to monitor the ablation status of the lesion throughout the entire ablation process, especially for thyroid nodule ablation, the nodule contour is unclear, resulting in ultrasound being unable to determine whether the ablation range completely covers the tumor and whether the ablation safety boundary has been reached.

[0004] At present, doctors rely more on naked-eye observation and clinical experience to complete this step of the work, which may result in large errors and make precise ablation impossible.

[0005] Therefore, it is crucial to develop a standardized evaluation system for real-time intraoperative positioning and monitoring of lesions and accurate postoperative evaluation of ablation effects in thyroid ablation procedures. Summary of the invention

[0006] The purpose of the present invention is to address the monitoring problem in thyroid ablation procedures and propose a thyroid ablation intraoperative monitoring method, system, storage medium and electronic device. By performing organ segmentation on thyroid imaging data where lesions can be clearly seen before surgery, marking lesions, determining the ablation range and safety boundaries, and achieving real-time and rapid registration of ultrasound images during ablation with preoperative three-dimensional images, the changes in the lesion area during real-time ablation can be effectively observed, ensuring the ablation effect and reducing the recurrence rate.

[0007] The technical solution of the present invention is:

[0008] The present invention provides a thyroid ablation intraoperative monitoring system, comprising:

[0009] A data acquisition unit, used for acquiring preoperative imaging volume data of the thyroid organ;

[0010] An image segmentation unit, configured to perform organ segmentation on the preoperative image volume data, obtain thyroid organ segmentation regions and lesion regions, and generate an ablation safety region based on spatial information of the lesion region;

[0011] A three-dimensional reconstruction unit is used to perform three-dimensional reconstruction and visualization according to the segmented area, lesion area and ablation safety area of ​​the thyroid organ, and obtain reconstructed organ volume data;

[0012] The dynamic monitoring unit is used to perform real-time registration of the preoperative image volume data with the ultrasound image sequence obtained during the operation to observe the location of the lesion, obtain the thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of the ablation.

[0013] Furthermore, in the data acquisition unit, acquiring preoperative image volume data of the thyroid organ includes:

[0014] Receiving real-time ultrasound image data, wherein the ultrasound image data is a two-dimensional ultrasound image sequence with spatial coordinate information obtained by continuously scanning the entire coverage area of ​​the thyroid gland along a preset path through an electromagnetic sensor bound to the ultrasound probe;

[0015] Find the spatial information of the two-dimensional ultrasound image sequence to determine the boundary of the sequence, and construct the pixel points within the boundary through the spatial interpolation algorithm to form the preoperative image volume data;

[0016] Among them, the preset path is pushed along the sagittal plane or the transverse plane, or a sector scan is performed at a fixed position.

[0017] Furthermore, in the data acquisition unit, obtaining preoperative image volume data of the thyroid organ includes:

[0018] Importing existing medical image data, wherein the type of the medical image data includes at least one of an ultrasound image, a CT image, an MRI image, and a PET image; the medical image data has multiple layers of slice images, and each slice image has position information in a three-dimensional space;

[0019] The slice images are combined to form preoperative image volume data.

[0020] Furthermore, in the image segmentation unit, performing organ segmentation on the preoperative image volume data to obtain thyroid organ segmentation regions and lesion regions includes:

[0021] Performing organ segmentation on the preoperative image volume data to obtain the thyroid organ segmentation area and the lesion area includes: automatically segmenting or manually outlining the thyroid organ based on AI to obtain the segmentation data of the thyroid organ, browsing from the cross-section, sagittal plane, and coronal plane to find the frame with the lesion area and outlining the lesion area.

[0022] Furthermore, in the image segmentation unit, generating an ablation safety area based on the spatial information of the lesion area includes:

[0023] Traverse the lesion area and extract the lesion contour points in each frame;

[0024] Transform the lesion contour points into the physical space to obtain the lesion contour points in the physical space;

[0025] Search for adjacent points around the lesion contour point in the physical space, fit the surface, and record the normal vector of the point;

[0026] The physical space lesion contour point is moved by a preset unit along the normal vector direction to obtain the corresponding safe area contour point;

[0027] Traverse the lesion contour points in the physical space, obtain the safe area contour points extended by each contour point, and form an ablation safe area.

[0028] Furthermore, the 3D reconstruction unit specifically performs the following steps:

[0029] Using the spatial information contained in the nifty format data, the physical coordinates of the surface vertices of the segmented data are calculated through the Marchingcube algorithm to form a reconstructed organ mesh and obtain a 3D visualization image.

[0030] Furthermore, the dynamic monitoring unit includes:

[0031] An intraoperative ultrasound data processing module is used to obtain intraoperative ultrasound image data and segment it to obtain an ultrasound image sequence to be registered and an organ segmentation mask sequence;

[0032] 2.5D-3D elastic registration module, used to take the preoperative image volume data, thyroid organ segmentation area, ultrasound image sequence to be registered, and organ segmentation mask sequence as the input of the deformation registration algorithm, perform elastic registration once, obtain the rigid transformation matrix RT and the three-dimensional elastic deformation field DDF, and apply the three-dimensional elastic deformation field to the preoperative image volume data and ablation safety area to obtain the deformed volume data and safety area;

[0033] A real-time secondary registration module is used to perform secondary elastic registration at regular intervals during the ablation process, and to perform the deformation registration algorithm in the 2.5D-3D elastic registration module on the cached intraoperative ultrasound image data, preoperative image volume data, and thyroid organ segmentation area to obtain a rigid transformation matrix RT´ and a three-dimensional elastic deformation field DDF´, and to apply the three-dimensional elastic deformation field to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and safety area;

[0034] The ablation completion confirmation module is used to fuse the real-time ultrasound image with the deformed volume data and the safe area, and confirm the completion of the ablation of the lesion and the safe area based on the fusion result.

[0035] Furthermore, the dynamic monitoring unit also includes:

[0036] The post-contrast re-registration module is used to perform re-registration by the 2.5D-3D elastic registration module and the real-time secondary registration module after the injection of contrast agent is completed, if there is still an unablated part in the ultrasound contrast image, until the ablation is completed.

[0037] Furthermore, in the 2.5D-3D elastic registration module, the three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area, including:

[0038] First, the original data is rigidly rotated and translated, and then the elastic deformation field is used to calculate the position of each point after the change, and finally the pixel value of each position in the volume data and the safe area is recalculated.

[0039] Furthermore, in the real-time secondary registration module, secondary elastic registration is performed periodically, including:

[0040] A queue-attributed cache is set to save n frames of thyroid ultrasound image data and mask data for thyroid contour segmentation. When the data in the cache is greater than n frames, the first frame is deleted, the latest frame is added to the last position, and the data in the cache and the preoperative volume data are deformed and registered.

[0041] A method for monitoring during thyroid ablation surgery, based on the system, comprises:

[0042] S1. Obtain preoperative imaging data of the thyroid organ;

[0043] S2, performing organ segmentation on the preoperative image volume data to obtain thyroid organ segmentation regions and lesion regions, and generating ablation safety regions based on spatial information of the lesion regions;

[0044] S3, perform three-dimensional reconstruction and visualization based on the segmented area of ​​the thyroid organ, the lesion area, and the ablation safety area to obtain the reconstructed organ volume data;

[0045] S4. Perform real-time registration of the preoperative image volume data with the ultrasound image obtained during the operation to observe the location of the lesion; and perform elastic registration based on the intraoperative image sequence, organ segmentation mask sequence and preoperative image volume data to obtain the thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of ablation.

[0046] A computer-readable storage medium stores a computer program, which, when executed, implements the thyroid ablation intraoperative monitoring method.

[0047] An electronic device comprises: one or more processors; the processor is used to store one or more programs; when the one or more programs are executed by the one or more processors, the thyroid ablation intraoperative monitoring method is implemented.

[0048] Beneficial effects of the present invention:

[0049] The present invention eliminates the need for 3D reconstruction of data collected during surgery by performing one-time registration plus multiple fast secondary precise registrations, making it easier to monitor the deformation of the thyroid gland during ablation. The intraoperative and preoperative registrations solve the problem of unclear positioning of ultrasonic lesions during surgery, thereby obtaining better ablation results.

[0050] The thyroid ablation intraoperative monitoring system disclosed in the present invention acquires preoperative ultrasound image volume data, segments and reconstructs the thyroid organ, lesions and ablation safety area in three dimensions, and realizes real-time intraoperative monitoring; specifically includes: using electromagnetic positioning technology to collect two-dimensional ultrasound sequences with spatial information or importing existing medical image data; generating ablation safety areas based on lesion contour points; and using Marching cube algorithm for three-dimensional reconstruction and visualization. In the dynamic monitoring unit, during the ablation process, the preoperative volume data is elastically registered with the intraoperative real-time ultrasound image to dynamically monitor thyroid deformation and lesion ablation status; by setting a cache queue, secondary fine registration is performed regularly to improve registration accuracy. Finally, contrast agent is injected for re-registration to confirm that the ablation is completed; real-time and accurate monitoring of the entire process of thyroid minimally invasive ablation surgery is achieved, the safety and effectiveness of the surgery is improved, and an important auxiliary decision-making basis is provided for clinicians.

[0051] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0053] Figure 1 A flow chart of intraoperative monitoring during thyroid ablation according to an embodiment of the present invention is shown.

[0054] Figure 2 A schematic diagram of a real-time ultrasound image data reconstruction process according to an embodiment of the present invention is shown.

[0055] Figure 3 A schematic diagram of organ segmentation of preoperative image volume data according to an embodiment of the present invention is shown.

[0056] Figure 4A schematic diagram showing the annotation of lesion regions in preoperative image volume data and the generation of ablation safety regions according to an embodiment of the present invention is shown.

[0057] Figure 5 A schematic diagram showing a schematic diagram of preoperative organ and lesion area reconstruction according to an embodiment of the present invention.

[0058] Figure 6 A dynamic monitoring flow chart according to an embodiment of the present invention is shown.

[0059] Figure 7 A schematic diagram of an intraoperative initial elastic registration according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0060] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0061] Example 1

[0062] The present invention provides a thyroid ablation intraoperative monitoring system, comprising:

[0063] A data acquisition unit, used for acquiring preoperative imaging volume data of the thyroid organ;

[0064] An image segmentation unit, configured to perform organ segmentation on the preoperative image volume data, obtain a thyroid organ segmentation region and a lesion region, and generate an ablation safety region based on spatial information of the lesion region;

[0065] A three-dimensional reconstruction unit is used to perform three-dimensional reconstruction and visualization according to the segmented area, lesion area and ablation safety area of ​​the thyroid organ, and obtain reconstructed organ volume data;

[0066] The dynamic monitoring unit is used to perform real-time registration of the preoperative image volume data with the ultrasound image sequence obtained during the operation to observe the location of the lesion, obtain the thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of the ablation.

[0067] In one example, in the data acquisition unit, obtaining the preoperative image volume data of the thyroid organ includes: online acquisition of ultrasound images or importing existing data:

[0068] Online acquisition of ultrasound images includes: receiving real-time ultrasound image data, which is acquired by an electromagnetic sensor bound to an ultrasound probe, and obtaining a two-dimensional ultrasound image sequence with spatial coordinate information by continuously scanning the entire coverage area of ​​the thyroid gland along a preset path; finding the spatial information of the two-dimensional ultrasound image sequence to determine the boundary of the sequence, and constructing pixel points within the boundary through a spatial interpolation algorithm to form preoperative image volume data.

[0069] Importing existing data includes: importing existing medical image data, the type of which includes at least one of ultrasound images, CT images, MRI images and PET images; the medical image data has multiple layers of slice images, each slice image has position information in three-dimensional space; and the slice images are combined to form preoperative image volume data.

[0070] In one example, in the image segmentation unit, performing organ segmentation on the preoperative image volume data to obtain a thyroid organ segmentation region and a lesion region includes:

[0071] Performing organ segmentation on the preoperative image volume data to obtain the thyroid organ segmentation area and the lesion area includes: automatically segmenting or manually outlining the thyroid organ based on AI to obtain the segmentation data of the thyroid organ, browsing from the cross-section, sagittal plane, and coronal plane to find the frame with the lesion area and outlining the lesion area.

[0072] Based on the spatial information of the lesion area, an ablation safe area is generated, including:

[0073] Traverse the lesion area and extract the lesion contour points in each frame;

[0074] Transform the lesion contour points into the physical space to obtain the lesion contour points in the physical space;

[0075] Search for adjacent points around the lesion contour point in the physical space, fit the surface, and record the normal vector of the point;

[0076] The physical space lesion contour point is moved by a preset unit along the normal vector direction to obtain the corresponding safe area contour point;

[0077] Traverse the lesion contour points in the physical space, obtain the safe area contour points extended by each contour point, and form an ablation safe area.

[0078] In one example, the 3D reconstruction unit specifically performs the following steps:

[0079] Using the spatial information contained in the nifty format data, the physical coordinates of the surface vertices of the segmented data are calculated through the Marchingcube algorithm to form a reconstructed organ mesh and obtain a 3D visualization image.

[0080] In one example, the dynamic monitoring unit includes:

[0081] An intraoperative ultrasound data processing module is used to obtain intraoperative ultrasound image data and segment it to obtain an ultrasound image sequence to be registered and an organ segmentation mask sequence;

[0082] 2.5D-3D elastic registration module, used to take the preoperative image volume data, thyroid organ segmentation area, ultrasound image sequence to be registered, and organ segmentation mask sequence as the input of the deformation registration algorithm, perform elastic registration once, obtain the rigid transformation matrix RT and the three-dimensional elastic deformation field DDF, and apply the three-dimensional elastic deformation field to the preoperative image volume data and ablation safety area to obtain the deformed volume data and safety area;

[0083] A real-time secondary registration module is used to perform secondary elastic registration at regular intervals during the ablation process, and to perform the deformation registration algorithm in the 2.5D-3D elastic registration module on the cached intraoperative ultrasound image data, preoperative image volume data, and thyroid organ segmentation area to obtain a rigid transformation matrix RT´ and a three-dimensional elastic deformation field DDF´, and to apply the three-dimensional elastic deformation field to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and safety area;

[0084] The ablation completion confirmation module is used to fuse the real-time ultrasound image with the deformed volume data and the safe area, and confirm the completion of the ablation of the lesion and the safe area based on the fusion result.

[0085] In one example, the dynamic monitoring unit further includes:

[0086] The post-contrast re-registration module is used to perform re-registration by the 2.5D-3D elastic registration module and the real-time secondary registration module after the injection of contrast agent is completed, if there is still an unablated part in the ultrasound contrast image, until the ablation is completed.

[0087] In one example, in the 2.5D-3D elastic registration module, a three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area, including:

[0088] First, the original data is rigidly rotated and translated, and then the elastic deformation field is used to calculate the position of each point after the change, and finally the pixel value of each position in the volume data and the safe area is recalculated.

[0089] In the real-time secondary registration module, secondary elastic registration is performed regularly, including:

[0090] A queue-attributed cache is set to save n frames of thyroid ultrasound image data and mask data for thyroid contour segmentation. When the data in the cache is greater than n frames, the first frame is deleted, the latest frame is added to the last position, and the data in the cache and the preoperative volume data are deformed and registered.

[0091] Example 2

[0092] The present invention provides a thyroid ablation intraoperative monitoring method, based on the above system, comprising:

[0093] S1. Obtain preoperative imaging data of the thyroid organ;

[0094] S2, performing organ segmentation on the preoperative image volume data to obtain thyroid organ segmentation regions and lesion regions, and generating ablation safety regions based on spatial information of the lesion regions;

[0095] S3, perform three-dimensional reconstruction and visualization based on the segmented area of ​​the thyroid organ, the lesion area, and the ablation safety area to obtain the reconstructed organ volume data;

[0096] S4. Perform real-time registration of the preoperative image volume data with the ultrasound image obtained during the operation to observe the location of the lesion; and perform elastic registration based on the intraoperative image sequence, organ segmentation mask sequence and preoperative image volume data to obtain the thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of ablation.

[0097] The monitoring process during thyroid ablation of the present invention is as follows Figure 1 As shown, specifically including;

[0098] S1. Online acquisition of ultrasound images or import of existing data;

[0099] S1.1. Use FreeHand to collect ultrasound images online;

[0100] The sensor of the electromagnetic positioning system is fixed to the probe in the form of a connector. The thyroid gland is continuously scanned at a speed of not less than 40 frames per second along a fixed path (for example, along the sagittal plane or transverse section, or by sector scanning at a fixed position). The scanning process covers the entire coverage area of ​​the thyroid gland, and a series of two-dimensional ultrasound sequence images with spatial coordinate information are obtained. , a single ultrasound image is saved in nifty format; then the sequence boundary is determined by finding the spatial information of the sequence to form a new volume data , each pixel value of the new volume data is obtained through the spatial interpolation algorithm, and the process of generating a new outer inclusion from the spatial sequence is as follows Figure 2 As shown, the new volume data is completed from the real-time ultrasound sequence acquired Schematic diagram of the reconstruction process.

[0101] S1.2, import existing medical data;

[0102] Use the image data import function to import existing image data. The data formats include dicom, nii, and mha. The data types include but are not limited to CT / MRI / US / PET, etc. These data formats save multi-layer slice images. Each image slice contains position information in three-dimensional space. These slice images are combined to form a three-dimensional volume data, referred to as , as the basis for reconstructing organ contours.

[0103] Steps S1.1 and S1.2 are two-or-the-other options in the system, and you can only choose one to import.

[0104] S2, organ segmentation and generation of safe area for lesion ablation;

[0105] Based on the preoperative image volume data obtained in step S1, the volume data or , perform automatic segmentation or manual delineation of the thyroid organ to obtain segmentation data of the thyroid organ , browse from the transverse, sagittal, and coronal planes to find the frame with the lesion area and outline the lesion area At the same time, based on the spatial information of the lesion delineation result, an ablation safety area is generated that wraps the lesion contour in space and does not exceed the organ contour. ; Thyroid organ segmentation area , lesion area and ablation safety zone They are all saved in nifty format, which contains the three-dimensional information of the volume data;

[0106] Typically, the spatial transformation matrix in nifty format is used , which represents the pixel-to-physical space transformation matrix in the volume data, calculates the coordinates of each pixel point in any volume data to the real physical space based on the following formula.

[0107]

[0108] Next, based on volume data Generate ablation safe zone volume data , the specific method is as follows:

[0109] S2.1. Usually, there is still a positive result about 5-10mm outside the lesion boundary seen on the MRI. In order to achieve a better treatment effect on the lesion area, a safe ablation area is planned. This area is larger than the lesion area but does not exceed the organ itself, which can achieve excellent treatment effects.

[0110] S2.2. Traversing the lesion area , extract the contour points of the frame with lesion annotation, recorded as lesion contour points , for the lesion contour point All points in are transformed into physical space using Formula 1 to obtain the physical space lesion contour points , there are physical space lesion contour points Any point in , represents the jth point in layer i. Search for its 2k+1 adjacent points in the adjacent 2n+1 planes, and use this batch of points to fit the surface , let the normal vector of this point be ;

[0111] S2.3, at this time along Direction of movement Units, the corresponding safety zone contour points are obtained according to the following formula , ,

[0112]

[0113] Get all the points of the expanded safe area, select the same frame number, fit the new contour for the points with the same frame number, and fill the contour to form a mask, so as to form the volume of the new safe area expansion ; The schematic diagram of organ segmentation, lesion annotation and safe area generation of the entire volume data is shown in Figure 3 , Figure 4 shown.

[0114] S3, 3D reconstruction and visualization

[0115] The segmented region of the thyroid organ with annotation information generated based on steps S1 and S2 , lesion area and ablation safety zone , complete three-dimensional reconstruction and 3D visualization of organs and lesion areas.

[0116] The specific implementation steps are as follows: Based on nifty data The spatial information contained in , using the following formula:

[0117] ;

[0118] in: express The actual position of each pixel in the physical coordinate system, express The coordinates of each pixel in the image coordinate system. The physical coordinates of the surface vertices of the segmented data are calculated by the Marchingcube algorithm to form a reconstructed organ mesh for three-dimensional visualization. The reconstructed schematic diagram is shown in Figure 5 shown.

[0119] S4, dynamically monitor the ablation process of the lesion area and whether the ablation is completed;

[0120] After completing the three-dimensional reconstruction, the intraoperative real-time registration process is started to perform real-time registration on the volume data reconstructed online before the operation or imported from the file and the intraoperative ultrasound to facilitate the observation of the location of the lesion. The present invention adopts a 2.5D to 3D registration method that combines organ morphology and image information, and uses intraoperative image sequences and mask sequences, plus preoperative volume data, to achieve flexible registration. The 2.5D to 3D representation input is a multi-sequence 2D image data and a 3D volume data.

[0121] The advantage of this method is that it does not require multiple 2D sequences to reconstruct a 3D volume data for 3D-3D registration, which saves a lot of time. After reconstruction, multiple images are collected for the first time to complete the registration. In the subsequent process, multiple sequences can be cached to quickly perform secondary precise registration, while compensating for the deformation during the whole process and the mismatch caused by the patient's small-range movement. The entire dynamic monitoring flow chart is as follows Figure 6 The detailed process is as follows:

[0122] S4.1 obtains a set of ultrasound thyroid images, which can be obtained by using sector scanning, plain scanning, etc., similar to the process mentioned in S1.1, and performs real-time segmentation on each frame of the image to obtain the intraoperative ultrasound image sequence to be registered And the sequence of organ segmentation masks for each image ;

[0123] S4.2 First intraoperative registration: The real-time ultrasound image volume data obtained in S1 Or medical ultrasound image volume data , and the thyroid organ segmentation area , combined with the ultrasound image sequence to be registered in S4.1 and organ segmentation mask sequence As the input of the deformable registration algorithm, the rigid transformation matrices R and T of the preoperative volume data are obtained after registration. and Image and shape information are corrected or The 3D elastic deformation field DDF_In_Surgury (x, y, z) records the movement distance of each point in three directions after elastic deformation. DDF_In_Surgury (x, y, z) is applied to the volume data using the following formula;

[0124] First, the original volume is rigidly rotated and translated. Then, the position of each point after the change is calculated using the elastic deformation field. Then, the pixel value of each position of the volume data is recalculated to obtain the volume after the deformation of the preoperative volume data. , or , transform to The whole process is as follows Figure 7 Shown is a schematic diagram of the initial elastic registration during surgery;

[0125] =transform( ,DDF_in_Surgury(x,y,z));

[0126] S4.3 Perform secondary precise registration at regular intervals during surgery: After the first registration is completed, thyroid ablation begins. During the ablation process, the system sets a queue attribute cache to save n frames of thyroid ultrasound image data and mask data for thyroid contour segmentation. When the data in the cache is greater than n frames, the first frame is deleted, the latest frame is added to the last position, and the data in the cache is recorded as and , and is called periodically during the ablation process and and or , execute the deformation registration process in S4.2, regularly perform deformation registration on the preoperative image and the intraoperative thyroid image, monitor the thyroid deformation during the ablation process in real time, and provide guidance on the lesion location and monitoring of the ablation status. The intraoperative timed secondary precise registration is to save the first acquisition process in the cache through the caching method, and then perform the first registration, and clear the cache at the same time;

[0127] S4.4 After the real-time image in S4.3 is registered with the preoperative image, and the real-time image shows that the tissue within the lesion and the ablation safety zone has been ablated, inject contrast agent to view the real-time image of the angiography, and perform the registration again according to the operations in steps S4.1 and S4.2. The preoperative image, intraoperative ultrasound image, and intraoperative ultrasound contrast image are all registered. If it is confirmed in the intraoperative ultrasound image and the intraoperative ultrasound contrast image that the lesion safety zone has been ablated, the entire ablation process is completed. If there is still a part that has not been ablated, continue with steps S4.1, S4.2, and S4.3 until the ablation is completely completed.

[0128] The method of one-time registration plus multiple precise registrations proposed in the present invention can quickly perform elastic registration during thyroid ablation, monitor changes during thyroid ablation, ensure a relatively efficient ablation range, and improve the success rate of thyroid ablation.

[0129] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A thyroid ablation intraoperative monitoring system, characterized in that: include: A data acquisition unit, used for acquiring preoperative imaging volume data of the thyroid organ; An image segmentation unit, configured to perform organ segmentation on the preoperative image volume data, obtain a thyroid organ segmentation region and a lesion region, and generate an ablation safety region based on spatial information of the lesion region; A three-dimensional reconstruction unit is used to perform three-dimensional reconstruction and visualization according to the segmented area, lesion area and ablation safety area of ​​the thyroid organ, and obtain reconstructed organ volume data; Dynamic monitoring unit, used to register preoperative image volume data with ultrasound image sequences acquired during surgery in real time to observe the location of the lesion, obtain thyroid deformation and lesion ablation status, and dynamically monitor the ablation process and completion status of the lesion area; The dynamic monitoring unit comprises: An intraoperative ultrasound data processing module is used to obtain intraoperative ultrasound image data and segment it to obtain an ultrasound image sequence to be registered and an organ segmentation mask sequence; 2.5D-3D elastic registration module, used to take the preoperative image volume data, thyroid organ segmentation area, ultrasound image sequence to be registered, and organ segmentation mask sequence as the input of the deformation registration algorithm, perform elastic registration once, obtain the rigid transformation matrix RT and the three-dimensional elastic deformation field DDF, and apply the three-dimensional elastic deformation field to the preoperative image volume data and ablation safety area to obtain the deformed volume data and safety area; A real-time secondary registration module is used to perform secondary elastic registration at regular intervals during the ablation process, and to perform the deformation registration algorithm in the 2.5D-3D elastic registration module on the cached intraoperative ultrasound image data, preoperative image volume data, and thyroid organ segmentation area to obtain a rigid transformation matrix RT´ and a three-dimensional elastic deformation field DDF´, and to apply the three-dimensional elastic deformation field to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and safety area; The ablation completion confirmation module is used to fuse the real-time ultrasound image with the deformed volume data and the safe area, and confirm the completion of the ablation of the lesion and the safe area based on the fusion result.

2. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that In the data acquisition unit, acquiring preoperative image volume data of the thyroid organ includes: Receiving real-time ultrasound image data, wherein the ultrasound image data is a two-dimensional ultrasound image sequence with spatial coordinate information obtained by continuously scanning the entire coverage area of ​​the thyroid gland along a preset path through an electromagnetic sensor bound to the ultrasound probe; The spatial information of the two-dimensional ultrasound image sequence is found to determine the boundary of the sequence, and the pixel points within the boundary are constructed through the spatial interpolation algorithm to form the preoperative image volume data.

3. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that In the data acquisition unit, obtaining preoperative image volume data of the thyroid organ includes: Importing existing medical image data, wherein the type of the medical image data includes at least one of an ultrasound image, a CT image, an MRI image, and a PET image; the medical image data has multiple layers of slice images, and each slice image has position information in a three-dimensional space; The slice images are combined to form preoperative image volume data.

4. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that In the image segmentation unit, performing organ segmentation on the preoperative image volume data to obtain thyroid organ segmentation regions and lesion regions includes: Perform automatic segmentation or manual outlining of the thyroid organ based on AI to obtain segmentation data of the thyroid organ, browse the cross-sectional, sagittal and coronal plane data, find the frame with the lesion area and outline the lesion area.

5. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that In the image segmentation unit, generating an ablation safety area based on spatial information of the lesion area includes: Traverse the lesion area and extract the lesion contour points in each frame; Transform the lesion contour points into the physical space to obtain the lesion contour points in the physical space; Search for adjacent points around the lesion contour point in the physical space, fit the surface, and record the normal vector of the point; The physical space lesion contour point is moved by a preset unit along the normal vector direction to obtain the corresponding safe area contour point; Traverse the lesion contour points in the physical space, obtain the safe area contour points extended by each contour point, and form an ablation safe area.

6. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that The three-dimensional reconstruction unit specifically performs the following steps: Using the spatial information contained in the nifty format data, the physical coordinates of the surface vertices of the segmented data are calculated through the Marchingcube algorithm to form a reconstructed organ mesh and obtain a 3D visualization image.

7. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that The dynamic monitoring unit also includes: The post-contrast re-registration module is used to perform re-registration by the 2.5D-3D elastic registration module and the real-time secondary registration module after the injection of contrast agent is completed, if there is still an unablated part in the ultrasound contrast image, until the ablation is completed.

8. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that In the 2.5D-3D elastic registration module, the three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area, including: First, the original data is rigidly rotated and translated, and then the elastic deformation field is used to calculate the position of each point after the change, and finally the pixel value of each position in the volume data and the safe area is recalculated.

9. The thyroid ablation intraoperative monitoring system according to claim 1, characterized in that In the real-time secondary registration module, secondary elastic registration is performed regularly, including: A queue-attributed cache is set to save n frames of thyroid ultrasound image data and mask data for thyroid contour segmentation. When the data in the cache is greater than n frames, the first frame is deleted, the latest frame is added to the last position, and the data in the cache and the preoperative volume data are deformed and registered.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed, a method for monitoring during thyroid ablation is implemented, including: S1. Obtain preoperative imaging data of the thyroid organ; S2, performing organ segmentation on the preoperative image volume data to obtain thyroid organ segmentation regions and lesion regions, and generating ablation safety regions based on spatial information of the lesion regions; S3, perform three-dimensional reconstruction and visualization based on the segmented area of ​​the thyroid organ, the lesion area, and the ablation safety area to obtain the reconstructed organ volume data; S4, real-time registration of the preoperative image volume data with the ultrasound image obtained during the operation to observe the location of the lesion; elastic registration is performed based on the intraoperative image sequence, the organ segmentation mask sequence and the preoperative image volume data to obtain the thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of the ablation, specifically including the following steps: Acquire intraoperative ultrasound image data, and segment it to obtain an ultrasound image sequence to be registered and an organ segmentation mask sequence; The preoperative image volume data, the thyroid organ segmentation area, the ultrasound image sequence to be registered, and the organ segmentation mask sequence are used as the input of the deformable registration algorithm, and an elastic registration is performed to obtain the rigid transformation matrix RT and the three-dimensional elastic deformation field DDF. The three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area. During the ablation process, secondary elastic registration is performed regularly, and the intraoperative ultrasound image data in the cache is used with the preoperative image volume data and the segmented area of ​​the thyroid organ to perform the deformation registration algorithm in the 2.5D-3D elastic registration module to obtain the rigid transformation matrix RT´ and the three-dimensional elastic deformation field DDF´, and the three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area; The real-time ultrasound image is fused with the deformed volume data and the safe area, and the ablation of the lesion and the safe area is confirmed to be complete based on the fusion results.

11. An electronic device, characterized in that: include: one or more processors; The processor is used to store one or more programs; When the one or more programs are executed by the one or more processors, a method for monitoring during thyroid ablation is implemented, including: S1. Obtain preoperative imaging data of the thyroid organ; S2, performing organ segmentation on the preoperative image volume data to obtain thyroid organ segmentation regions and lesion regions, and generating ablation safety regions based on spatial information of the lesion regions; S3, perform three-dimensional reconstruction and visualization based on the segmented area of ​​the thyroid organ, the lesion area, and the ablation safety area to obtain the reconstructed organ volume data; S4, real-time registration of the preoperative image volume data with the ultrasound image obtained during the operation to observe the location of the lesion; elastic registration is performed based on the intraoperative image sequence, the organ segmentation mask sequence and the preoperative image volume data to obtain the thyroid deformation and lesion ablation status, and dynamically monitor the ablation process of the lesion area and the completion of the ablation, specifically including the following steps: Acquire intraoperative ultrasound image data, and segment it to obtain an ultrasound image sequence to be registered and an organ segmentation mask sequence; The preoperative image volume data, the thyroid organ segmentation area, the ultrasound image sequence to be registered, and the organ segmentation mask sequence are used as the input of the deformable registration algorithm, and an elastic registration is performed to obtain the rigid transformation matrix RT and the three-dimensional elastic deformation field DDF. The three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area. During the ablation process, secondary elastic registration is performed regularly, and the intraoperative ultrasound image data in the cache is used with the preoperative image volume data and the segmented area of ​​the thyroid organ to perform the deformation registration algorithm in the 2.5D-3D elastic registration module to obtain the rigid transformation matrix RT´ and the three-dimensional elastic deformation field DDF´, and the three-dimensional elastic deformation field is applied to the preoperative image volume data and the ablation safety area to obtain the deformed volume data and the safety area; The real-time ultrasound image is fused with the deformed volume data and the safe area, and the ablation of the lesion and the safe area is confirmed to be complete based on the fusion results.

Images