Multi-needle steep pulse ablation path planning method and system based on three-dimensional electric field coverage

Abstract

The invention provides a multi-needle steep pulse ablation path planning method and system based on three-dimensional electric field coverage, and the method comprises the steps: carrying out the integration and calibration of image data obtained by all image devices, segmenting an ablation region, constructing a three-dimensional model of the ablation region, and carrying out the multi-modal image fusion; a needle entry point and an end point are selected on the T or S or C surface of the three-dimensional model in a visual interaction mode so as to generate an initial needle position layout, then field energy calculation is carried out according to needle position parameters, a three-dimensional field energy distribution model is generated, and the three-dimensional model and the three-dimensional field energy distribution model of the ablation area are compared one by one through voxelization; calculating a field energy coverage rate; if the field energy coverage rate does not meet the set requirement, the needle position is adjusted in a visual interaction mode till the field energy coverage rate meets the set requirement. The method has the advantages that accurate coverage of a tumor ablation area is achieved, dependence on experience of doctors is reduced, operation safety and treatment effect are improved, medical cost is reduced, and hospital benefits are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical navigation, in particular to a visual multi-needle steep pulse ablation path planning method and system suitable for tumor ablation treatment, combined with medical image three-dimensional reconstruction, multi-needle field energy calculation and coverage analysis. BACKGROUND

[0002] In the field of tumor treatment, traditional multi-needle ablation treatment, as a common technical means, can effectively ablate tumor tissue to a certain extent, but in the actual operation process, the layout of the ablation needle often mainly depends on the personal experience accumulated by the doctor for many years. The doctor roughly determines the insertion position and angle of the multi-needle ablation needle under the guidance of the image, in order to achieve the ablation coverage of the tumor tissue. However, this method lacks accurate quantitative evaluation of the field energy coverage range.

[0003] The currently applied path planning system still has obvious deficiencies in terms of advancement, and has not integrated the voxel analysis technology of three-dimensional field energy, and cannot perform comprehensive and refined grid analysis on the field energy distribution generated during the ablation process, so it cannot accurately determine whether the tumor tissue has been completely covered by the field energy, which has certain uncertainty for the treatment effect.

[0004] In addition, the existing single-needle field energy model is also relatively simple, which only simply linearly superimposes the field energy generated by each ablation needle, without fully considering the influence of tissue heterogeneity factors on the field energy distribution. For example, the existence of blood vessels is an important influencing factor. The blood in the blood vessels has strong heat conduction and flow, when the heat generated by ablation is transmitted to the vicinity of the blood vessels, the blood will quickly take away the heat, causing local heat to be quickly taken away, resulting in uneven heat distribution; the heat conduction and flow of blood vessels cause local heat dissipation to be enhanced, making it difficult to effectively accumulate the temperature in the ablation area, thereby affecting the heat ablation effect. This attenuation phenomenon may cause uneven heat distribution in the ablation area in the actual tumor ablation process, thereby affecting the ablation effect of the tumor tissue, and even part of the tumor tissue may not be fully ablated.

[0005] Therefore, the existing technology still has many limitations in simulating and predicting the actual field energy distribution of multi-needle ablation, and needs to be further improved and perfected to improve the accuracy and effectiveness of multi-needle ablation treatment, and to provide more reliable treatment options for tumor patients. SUMMARY

[0006] The purpose of the present application is to provide a multi-needle steep pulse ablation path planning method and system based on three-dimensional electric field coverage, which can realize accurate coverage of the tumor ablation area, reduce the dependence on the experience of doctors, and improve the safety of surgery and treatment effect.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solution: The multi-needle steep pulse ablation path planning method based on three-dimensional electric field coverage integrates, calibrates, and segments the ablation area from image data acquired by various imaging devices, constructs a three-dimensional model of the ablation area, and generates an initial needle position layout by selecting the entry point and endpoint on the T, S, or C planes of the three-dimensional model through a visual interactive method. Then, the field energy is calculated based on the needle position parameters to generate a three-dimensional field energy distribution model. The three-dimensional model of the ablation area and the three-dimensional field energy distribution model are then compared one by one through voxelization to calculate the field energy coverage. If the field energy coverage does not meet the set requirements, the needle positions are adjusted through a visual interactive method until the field energy coverage meets the set requirements.

[0008] I. Intraoperative Image Processing Intraoperative imaging equipment includes: (1) CT equipment: used to acquire high-resolution tomographic images and provide detailed anatomical information of the surgical area; (2) MRI equipment: used to acquire high-contrast soft tissue images and help distinguish between tumor tissue and healthy tissue; (3) Ultrasound equipment: used to acquire real-time dynamic ultrasound images of human soft tissues and organs and provide information such as tissue structure morphology, hemodynamics and lesion characteristics.

[0009] First, the data acquisition and transmission of each imaging device includes: (1) Real-time data acquisition: Each imaging device works synchronously to acquire the imaging data of the surgical area in real time. Let the ultrasound imaging data be... CT image data is MRI imaging data are , t (2) Data transmission and integration: The data acquired by each imaging device is transmitted to the computing processing unit through network communication. The transmitted image data can be represented as: ,in, In time t A collection of all image data acquired at any given time.

[0010] Secondly, each imaging device is synchronized and calibrated, specifically including: (1) Event synchronization: Each imaging device achieves time synchronization through a synchronization signal to ensure that the image data acquired at the same time point has time consistency. The formula for time synchronization is expressed as: ,in, , , (1) Timestamps for ultrasound, CT, and MRI image data, respectively; (2) Spatial calibration: Using a spatial calibration algorithm, spatial calibration is performed on the image data acquired by each imaging device to ensure that each image data is aligned in the same spatial coordinate system. The formula is as follows: ,in, For calibrated image data, For spatial calibration function, These are spatial coordinates.

[0011] Next, the image data is preprocessed, specifically including: (1) Image enhancement: The acquired raw image data is enhanced to improve the contrast and clarity of the image. The formula is as follows: ,in, For the enhanced image data, (2) Noise Removal: The noise removal algorithm is used to eliminate noise in the image and improve the image quality. The formula is as follows: ,in, The image data after denoising. This is the noise removal function.

[0012] Next, the ablation region is accurately identified and segmented from the preprocessed image data. Convolutional Neural Network (CNN) image segmentation algorithms are usually used to ensure the accuracy and reliability of the segmentation results. Specifically, this includes: (1) Data input: Inputting the preprocessed image data. ,in, For time variables, representing image data collected at different time points; (2) Convolutional Neural Network (CNN) structure design: Design and train a convolutional neural network (CNN) suitable for ablation region segmentation. The network structure includes multiple convolutional layers, pooling layers and fully connected layers. The specific structure is as follows: (2.1) Convolutional layer: Extracts local features of the image. The formula is: ,in, For the output of the convolutional layer, For preprocessed image data in location and time t pixel values, i,j The index of the convolution kernel. For convolution kernel weights, b For bias; (2.2) Pooling layer: reduces the dimension of the feature map and reduces the amount of computation, the formula is: ,in, This is the output of the pooling layer. For the output of the convolutional layer at position The pixel value, where max is the max pooling operation; (2.3) Fully Connected Layer: Features extracted from both convolutional and pooling layers are combined for classification, using the following formula: ,in, For the output of the fully connected layer, This is the output of the pooling layer. This is the weight matrix of the fully connected layer. This is the bias vector of the fully connected layer. (3) Segmentation process: The preprocessed image data is used as the activation function; The input is fed into a trained convolutional neural network, which will automatically perform feature extraction and region segmentation, and output the ablation region. The formula is as follows: ,in, In time t (3) Segmentation results at time step; (4) Post-processing of segmentation results: processing of segmentation results Post-processing, including morphological processing and edge correction, is performed to improve the accuracy of the segmentation results; the formula for morphological processing is as follows: ,in, The segmentation result is after morphological processing. Morph is the morphological processing function, which includes erosion and dilation. The formula for edge correction is: ,in, This represents the segmentation result after edge correction, where Edge is the edge detection and correction function.

[0013] Finally, a 3D reconstruction of the ablation area is performed to generate a 3D model of the ablation area. The purpose of the 3D reconstruction is to provide a three-dimensional view of the ablation area during the procedure, allowing for a better understanding of the ablation process. The 3D reconstruction calculation is based on volume integral rendering, and the formula is as follows: ,in, The image represents the reconstructed 3D image, where V is the reconstructed volume and dV is a volume element. In time t The ablation region after segmentation and edge correction. The ablation region obtained after segmentation and edge correction... Input the 3D reconstruction module, and use the volume integration method to... Integrating within volume V generates a three-dimensional ablation region image. The generated 3D image Output, used for subsequent processing and display.

[0014] II. Multimodal Image Fusion The composite image is generated by fusing image data from different modalities, as shown in the following formula: ,in, In time t real-time ultrasound images, In time t CT images at different times, In time t MRI images at different times, , , To reconstruct the volume weighting coefficients, satisfying .

[0015] The weighting coefficients are determined based on clinical needs and imaging characteristics. , , The value is used to calculate the image data of different modalities according to the fusion formula to generate a composite image. It is then output for subsequent processing and display.

[0016] III. Multi-needle path planning (1) Define a three-dimensional coordinate system Establishing a 3D spatial coordinate system based on the DICOM coordinate system of multimodal image fusion : T-plane (cross section): A horizontal slice perpendicular to the Z-axis (XY plane); S-plane (sagittal plane): a longitudinal slice perpendicular to the X-axis (YZ plane); C-plane (coronal plane): A longitudinal slice perpendicular to the Y-axis (XZ plane).

[0017] (2) Interactive planning Select a marker point: Select marker point M on the T-plane. This marker point is a pre-designed auxiliary position for the patient before taking a CT or MRI. Select start and end points: Click the needle entry point on the T-plane. and the finish line The system automatically connects the needle entry point and the needle endpoint to generate a needle combination. At the same time, the system automatically displays the corresponding projections on two other planes simultaneously. Movement and calibration: The needle combination can be moved in any plane until the specific ablation position is found and the needle length L (mm) is displayed in real time. Coordinate transformation: Converting two-dimensional planar coordinates to three-dimensional spatial coordinates, selecting points on the T-plane. Corresponding three-dimensional coordinates ,in The slice number of the current T-side is multiplied by the layer thickness (e.g., 1 mm).

[0018] (3) Multi-needle synergistic layout Supports simultaneous planning of multiple needles (2-10 needles): The system automatically detects whether the needle spacing is ≥ the safety threshold (25mm), and prompts for adjustment if there is a conflict; Dynamic safety boundary: Distinguish vascular / nerve regions based on CT values, and prohibit needle tracts from passing through high-risk areas (bones or blood vessels with CT values ​​≥100HU).

[0019] (4) Needle position parameter generation The needle position parameters are transmitted to the field energy calculation module to generate the superposition effect of multiple needle electric fields within a region: Direction vector V originates from the endpoint and needle insertion point The difference is determined by the following formula: ; Depth d is the magnitude of the direction vector V, and is calculated using the following formula: Angle of elevation It is the angle of the direction vector in the vertical direction, and the calculation formula is: ; Azimuth It is the angle of the direction vector in the horizontal direction, and the calculation formula is: .

[0020] (5) Real-time feedback and dynamic adjustment When dragging the needle track endpoint in the visual interface, the system updates the needle position parameters in real time and triggers the following process through an event-driven mechanism: Collision detection: Check whether the new needle path intersects with existing needle paths or danger zones; Parameter recalculation: update direction vector, angle, and depth; Field energy regeneration: Call the field energy calculation module to re-solve the electric field distribution based on the new needle position (GPU accelerated, time ≤60s), and time ≤20s for each additional needle; Coverage refresh: The coverage analysis module outputs updated parameters such as coverage, tumor volume, field energy volume, and longest diameter of the tumor, and the visualization interface simultaneously renders a heat map.

[0021] IV. Field Energy Calculation The needle positions are arranged in groups of two needles. The space charge distribution density of each group of needles is calculated using Coulomb's law, and then the space electric field energy is calculated using Maxwell's equations. Finally, a three-dimensional field energy distribution model is constructed using visualization technology.

[0022] V. Field Energy Coverage Analysis (1) Model acquisition and validation Based on the DICOM coordinate system of multimodal image fusion, a three-dimensional spatial coordinate system is used to obtain the automatically segmented ablation region STL (or VTP) model file and the ablation region STL (or VTP) model file generated by the field energy calculation module. The model file is loaded through the three-dimensional model reading module to perform normal direction correction and closure verification.

[0023] (2) Spatial mapping A voxel space is constructed based on multimodal image parameters, and the model file is mapped to a binary voxel matrix, outputting a binary voxel matrix (1=inside the model, 0=outside).

[0024] (3) Coverage calculation The formula for calculating the field energy coverage rate is: Where Cabla represents the field energy coverage, TumorVoxels represents the ablation region after voxelization, and AblationVoxels represents the field energy region after voxelization. The safety margin weight is used for marginCoverage, which represents the indication coverage rate.

[0025] A multi-needle steep pulse ablation path planning system based on three-dimensional electric field coverage includes an intraoperative image processing module, a multi-needle path planning module, a field energy calculation module, a coverage analysis module, a visualization interaction module, and a high-voltage steep pulse generator. The intraoperative image processing module imports and integrates image data from several devices, then segments the ablation area and performs three-dimensional reconstruction to generate a three-dimensional model of the ablation area. The multi-needle path planning module can select the entry point and endpoint on the T, S, or C planes of the three-dimensional model to generate an initial needle position layout and can adjust the needle position distribution. The field energy calculation module calculates the field energy based on the needle position parameters and generates a three-dimensional field energy distribution model. The coverage analysis module acquires the three-dimensional model and the three-dimensional field energy distribution model of the ablation area, compares them one by one through voxelization, and calculates the field energy coverage rate. The visualization interaction module displays the needle position distribution, field energy thermal distribution, and field energy coverage rate, and can be combined with the multi-needle path planning module to adjust the needle position distribution in real time. The high-voltage steep pulse generator generates high-voltage steep pulses to ablate the ablation area according to the ablation path planning.

[0026] Compared with the prior art, the present invention has the following advantages: This invention presents a multi-needle steep pulse ablation path planning method and system based on three-dimensional electric field coverage. It solves the technical problem in traditional multi-needle ablation therapy where the placement of ablation needles relies heavily on the physician's personal experience and lacks quantitative assessment of the field energy coverage. This achieves precise coverage of the tumor ablation area, reduces reliance on physician experience, improves surgical safety and treatment efficacy, lowers medical costs, and enhances hospital efficiency. Specifically, by combining three-dimensional voxelized field energy calculation with a multi-needle electric field superposition model, the electric field distribution in the ablation area is precisely quantified, ensuring a tumor coverage rate ≥120%, significantly reducing the risk of tumor residue. Real-time interactive visualization and multi-plane collaborative needle placement reduce the average needle placement time from 45 minutes to within 10 minutes, shortening the operation time and reducing operational complexity. The field energy calculation results are updated within 60 seconds after needle position adjustment, achieving a closed-loop "planning-feedback" system. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0028] Figure 2 This is a schematic diagram of the system functions of the present invention. Detailed Implementation

[0029] The embodiments of the present invention will now be described in further detail with reference to the accompanying drawings. Example

[0030] A 55-year-old male patient with liver cancer had a tumor located in the right lobe of the liver, measuring approximately 3.2cm × 2.8cm × 3.0cm, with an irregular shape, and located near a branch of the portal vein. In traditional ablation therapy, due to the complex shape of the tumor and its proximity to blood vessels, relying solely on the doctor's experience to perform multiple needle placements is insufficient to ensure complete tumor coverage, and there is also a high risk of complications.

[0031] like Figure 1 , 2 As shown, the multi-needle steep pulse ablation path planning based on three-dimensional electric field coverage of this invention is implemented using the following specific steps: S1) Intraoperative Image Acquisition: In the operating room, an intraoperative CT scanner is first used to scan the patient's liver, acquiring high-resolution tomographic images. Subsequently, an intraoperative MRI scanner is used to acquire soft tissue images of the liver at a resolution of 0.5mm × 0.5mm × 0.5mm, providing a basis for accurately distinguishing the tumor from surrounding healthy tissue. The data acquired by the CT and MRI scanners are transmitted to the computing unit in real time via a high-speed network.

[0032] S2) Data preprocessing: Image enhancement processing is performed on the acquired image data. Histogram equalization algorithm is used to improve image contrast and make tumor boundaries clearer. Then, a noise removal algorithm based on wavelet transform is used to eliminate Gaussian noise in the image and improve image quality.

[0033] S3) Ablation region segmentation: The preprocessed image data is input into a pre-trained convolutional neural network (CNN), which includes 5 convolutional layers, 3 pooling layers and 2 fully connected layers. After the convolutional layers extract local features, the pooling layers reduce the feature dimensionality, and the fully connected layers perform comprehensive classification, the segmentation result of the ablation region is output. Then, the segmentation result is subjected to morphological processing (erosion followed by dilation) to remove small noise points and holes, and edge correction is performed to make the tumor boundary smoother and more accurate.

[0034] S4) 3D Reconstruction: Input the segmented and edge-corrected ablation area data into the 3D reconstruction module, and use the volume integral rendering algorithm to perform 3D reconstruction. By performing integral calculation on voxels in 3D space, a 3D stereoscopic image of the liver tumor is generated, which intuitively shows the location, size, shape and relationship of the tumor with surrounding tissues.

[0035] S5) Multimodal image fusion: Based on clinical needs, ultrasound images, CT images and MRI images are fused. The weighting coefficients for ultrasound images, CT images and MRI images are set to 0.3, 0.4 and 0.3 respectively. A comprehensive image is generated according to the fusion formula, which allows doctors to obtain multiple modal image information on the same image at the same time and better understand the overall picture of the tumor.

[0036] S6) Multi-needle path planning: On the fused multimodal images, a three-dimensional spatial coordinate system is established based on the DICOM coordinate system. Pre-marked reference points are selected in the transverse section (T-plane), and the needle entry point and endpoint are confirmed by clicking. The system automatically generates a needle combination and displays the corresponding projections simultaneously in the sagittal (S-plane) and coronal (C-plane). Based on the shape and location of the tumor, the path of 5 ablation needles is planned. The system automatically detects that the needle spacing is greater than or equal to the safety threshold of 25mm. At the same time, high-risk areas such as blood vessels and nerves are avoided based on CT values ​​(bone CT value ≥100HU). The direction vector, depth, elevation angle, and azimuth angle of each needle are calculated to provide data support for subsequent field energy calculations.

[0037] S7) Field Energy Calculation and Coverage Analysis: The needle position parameters are transmitted to the field energy calculation module. Based on Maxwell's equations and Coulomb's law, the three-dimensional field energy distribution after the superposition of multiple needle electric fields is calculated. The field energy distribution is voxelized, and each voxel stores the electric field intensity value. Then, it is compared with the tumor model. The intersection of the voxel matrix of the ablation area and the field energy area is calculated through NumPy logic operations. The initial tumor coverage rate is 82%, which does not reach the expected 120% or more.

[0038] S8) Needle Position Adjustment and Real-time Feedback: Based on the coverage analysis results, the doctor can drag some needle path endpoints in the visualization interface to adjust the needle position layout; the system updates the needle position parameters in real time, automatically performs collision detection, parameter recalculation and field energy regeneration. After 3 adjustments, the field energy calculation results show that the tumor coverage rate has increased to 120%, meeting the treatment requirements; at the same time, the visualization interface renders the updated field energy heat map, which intuitively shows the field energy coverage range and intensity distribution.

[0039] S9) Ablation therapy execution: The final determined needle position parameters are transmitted to the high-voltage steep pulse generator, and ablation parameters such as pulse voltage, pulse width, and pulse frequency are set; the high-voltage steep pulse generator generates high-voltage steep pulses according to the set parameters, and transmits the pulses to the tumor area through the ablation needle to perform ablation therapy on the tumor; during the treatment process, the temperature and field energy changes of the ablation area are monitored in real time to ensure the safety and effectiveness of the treatment.

[0040] Through the ablation pathway planning and ablation treatment described above, the patient's liver tumor was effectively ablated, and no signs of tumor recurrence were observed during the 6-month follow-up. Compared with traditional ablation treatment, this system significantly improves the coverage and precision of tumor ablation, shortens the operation time, reduces reliance on the doctor's experience, and also reduces the risk of complications, providing a more precise, safe, and effective solution for the treatment of liver cancer patients.

[0041] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements and modifications without departing from the concept of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-needle steep pulse ablation path planning method based on three-dimensional electric field coverage, characterized in that: The image data acquired by various imaging devices are integrated, calibrated, and segmented into ablation areas. A three-dimensional model of the ablation area is constructed, and multimodal image fusion is performed. The initial needle position layout is generated by selecting the entry point and endpoint on the T, S, or C plane of the three-dimensional model through a visual interactive method. Then, the field energy is calculated based on the needle position parameters, and a three-dimensional field energy distribution model is generated. The three-dimensional model of the ablation area and the three-dimensional field energy distribution model are compared one by one through voxelization to calculate the field energy coverage. If the field energy coverage does not meet the set requirements, the needle position is adjusted through a visual interactive method until the field energy coverage meets the set requirements.

2. The multi-needle steep pulse ablation path planning method based on three-dimensional electric field coverage according to claim 1, characterized in that: When laying out the needles, multi-needle planning can be performed, and the system automatically detects whether the needle spacing and needle placement are safe.

3. The multi-needle steep pulse ablation path planning method based on three-dimensional electric field coverage according to claim 1, characterized in that: The needle positions are arranged in groups of two needles. The space charge distribution density of each group of needles is calculated using Coulomb's law, and then the space electric field energy is calculated using Maxwell's equations. Finally, a three-dimensional field energy distribution model is constructed using visualization technology.

4. The multi-needle steep pulse ablation path planning method based on three-dimensional electric field coverage according to claim 1, characterized in that: The formula for calculating the field energy coverage rate is: ;in, For field energy coverage, This refers to the ablation region after voxelization. This represents the field energy region after voxelization. For safety boundary weights, For indication coverage.

5. A multi-needle steep pulse ablation path planning system based on three-dimensional electric field coverage, characterized in that: The system includes an intraoperative image processing module, a multi-needle path planning module, a field energy calculation module, a coverage analysis module, a visualization interaction module, and a high-pressure steep pulse generator. The intraoperative image processing module imports image data from several devices, integrates and calibrates it, then segments the ablation area and performs 3D reconstruction to generate a 3D model of the ablation area, and performs multimodal image fusion. The multi-needle path planning module allows selection of the entry and exit points on the T, S, or C planes of the 3D model to generate an initial needle layout, and allows adjustment of the needle distribution. The field energy calculation module calculates the field energy based on the needle position parameters and generates a 3D field energy distribution model. The coverage analysis module acquires the 3D model and the 3D field energy distribution model of the ablation area, compares them one by one through voxelization, and calculates the field energy coverage. The visualization interaction module displays the needle distribution, field energy thermal distribution, and field energy coverage, and can be combined with the multi-needle path planning module to adjust the needle distribution in real time. The high-pressure steep pulse generator generates high-pressure steep pulses to ablate the ablation area according to the ablation path planning.

Images