Interventional surgical robot guided by ultrasonic imaging

Abstract

The invention discloses an interventional surgical robot guided by ultrasonic imaging, and belongs to the technical field of minimally invasive surgical robots. The interventional surgical robot comprises a planning module used for extracting tumor contour features, constructing an ablation energy field simulation model and automatically generating an optimal needle insertion path planning scheme; the auxiliary positioning module is used for acquiring a focus three-dimensional space initial position coordinate, establishing a focus positioning reference coordinate system and dynamically tracking the focus position in real time; the early warning module is used for establishing a three-dimensional model of a normal organ around the tumor, marking a dangerous area, optimizing an ablation moving path, calculating a real-time distance between a vaporization area and the dangerous area calibrated before an operation, comparing the real-time distance with a safe distance threshold value, and triggering an alarm strategy. Through cooperation of the three modules, a full-process closed loop of'preoperative planning-intraoperative positioning-danger early warning 'is constructed, a tumor contour and a simulated ablation energy field are extracted, dynamic tracking of small focuses is achieved, and complication risks are reduced through graded early warning linkage control.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of minimally invasive surgery robots, in particular to an ultrasound imaging guided interventional surgery robot. BACKGROUND

[0002] The ultrasound imaging guided interventional surgery robot is a minimally invasive surgery platform that deeply integrates real-time ultrasound images with mechanical arm control. It first uses an ultrasound probe to continuously collect high-frame-rate tomographic images on the patient's body surface or in the blood vessels, providing a "real-time map" for the surgical instrument, so that the surgeon or control system can see the relative position of the needle tip, catheter and surrounding blood vessels, and tumor at any time. Subsequently, the robot automatically or semi-automatically adjusts the advancement angle, depth and force of the instrument according to image feedback, enabling operations such as puncture, biopsy, catheterization or ablation, with an accuracy of typically millimeters. Compared with traditional manual operation, this robot eliminates hand tremor, overcomes errors caused by limited ultrasound viewing angle and respiratory motion, and ensures that the probe always maintains optimal contact with the tissue through force sensing and impedance control, significantly improving the precision and safety of interventional surgery.

[0003] However, the existing ultrasound imaging guided interventional surgery robot still has many technical bottlenecks to be improved in clinical application, for example: for small lesions with a diameter ≤5mm or lesions in remote locations, the lesion is easily lost during the operation due to factors such as patient respiratory motion and body position changes. The existing positioning method does not establish a cooperative mechanism for preoperative accurate calibration and intraoperative dynamic tracking, and cannot realize real-time correction of the lesion position, increasing the difficulty and missed treatment risk of ablation operation; and when ablation is performed near important organs such as intestines, trachea and large blood vessels, it is also difficult to monitor the distance between the ablation vaporization zone and the dangerous organs in real time; when the vaporization zone approaches the safety threshold, there is a lack of automatic warning, energy adjustment or operation pause linkage mechanism, which is prone to cause complications such as organ damage.

[0004] Therefore, the present application proposes an ultrasound imaging guided interventional surgery robot to solve the above problems. SUMMARY

[0005] To solve the above problems, the present application provides an ultrasound imaging guided interventional surgery robot, which constructs a "preoperative planning-intraoperative positioning-danger warning" whole-process closed loop, accurately extracts the tumor contour, simulates the ablation energy field, dynamically tracks small lesions, and reduces the risk of complications through hierarchical warning linkage control.

[0006] In order to achieve the above purpose, the technical scheme of the present application is as follows: an ultrasound imaging guided interventional surgery robot, comprising:

[0007] The planning module is configured to extract tumor contour features based on a three-dimensional ultrasound image of a tumor region of a patient, determine tumor feature information, and construct an ablation energy field simulation model, calculate energy field simulation results under different ablation parameters, and automatically generate an optimal needle insertion path planning scheme in combination with the energy field simulation results;

[0008] The auxiliary positioning module is configured to obtain a three-dimensional space initial position coordinate of a lesion based on an enhanced image of the lesion, record a relative position relationship between the lesion and surrounding landmark anatomical structures, and establish a lesion positioning reference coordinate system; the auxiliary positioning module is also configured to obtain an actual anatomical coordinate system of a patient during an operation, align the lesion positioning reference coordinate system established before the operation with the actual anatomical coordinate system of the patient during the operation, and track a lesion position in real time and dynamically;

[0009] The warning module is configured to establish a three-dimensional model of a normal organ surrounding a tumor, introduce the three-dimensional model of the normal organ into the ablation energy field simulation model, mark a dangerous area on the three-dimensional model of the normal organ, set a corresponding safety distance threshold, analyze a potential intersection between an ablation movement path generated by the ablation energy field simulation model and the dangerous area, and optimize the ablation movement path; the warning module is also configured to monitor a change in a vaporization zone of an ablation area in real time during the operation, extract vaporization zone contour and position information based on a vaporization zone segmentation model, calculate a real-time distance between the vaporization zone and a dangerous area labeled before the operation, compare the real-time distance with the safety distance threshold, and trigger a warning strategy.

[0010] Further, the planning module includes a preoperative image acquisition and processing unit, an ablation energy field simulation unit, and a needle insertion path planning unit;

[0011] The preoperative image acquisition and processing unit is configured to perform multi-dimensional scanning on a tumor region of a patient by using an ultrasound probe, acquire continuous tomographic ultrasound images, generate a three-dimensional ultrasound image of the tumor region by using a three-dimensional reconstruction algorithm, pre-process the three-dimensional ultrasound image of the tumor region, optimize feature differences between the tumor and surrounding normal tissues, and generate a first image; the preoperative image acquisition and processing unit is also configured to perform image segmentation by using a U-Net++ improved model based on the first image, and output three-dimensional coordinate information of the tumor to clearly indicate a size, a shape, and a spatial position relationship of the tumor with surrounding organ anatomical structures;

[0012] The ablation energy field simulation unit is configured to construct a multi-modal ablation energy field simulation system based on the first image, call a finite element analysis algorithm, calculate an energy diffusion process in the tumor and surrounding tissues based on a biological heat conduction equation, and generate an energy field distribution cloud image, a temperature change dynamic curve, and an effective ablation range;

[0013] The needle insertion path planning unit is configured to combine tumor contour data, energy field simulation results, and important anatomical structures labeled before the operation, and adopt an A *The path search algorithm automatically generates an optimal needle insertion path, and outputs the coordinates of the needle insertion point, the puncture angle and the pushing depth of the ablation needle.

[0014] Further, in the needle insertion path planning unit, the optimization basic strategy includes: the energy field completely covers the tumor, the path avoids dangerous structures and the puncture depth is the shortest.

[0015] Further, the needle insertion path planning unit is also used to support manual adjustment of the planning scheme of the optimal needle insertion path, and real-time update of the energy field simulation results and the risk assessment report after adjustment, to assist in confirming the final planning scheme.

[0016] Further, the auxiliary positioning module includes a preoperative contrast imaging and coordinate system establishment unit, an intraoperative image acquisition and registration unit, and a lesion shift monitoring and mechanical arm adjustment unit.

[0017] The preoperative contrast imaging and coordinate system establishment unit is used to preoperatively inject an ultrasound contrast agent through a vein, control an ultrasound probe to perform three-dimensional scanning on the lesion area after the contrast agent is enriched in the lesion area, collect enhanced ultrasound images of the lesion and surrounding tissues, and highlight the gray difference between lesions with a diameter ≤5mm and normal tissues; the enhanced ultrasound images are preprocessed, the three-dimensional contour of the small lesion is extracted through a U-Net++ improved model, the initial position coordinates of the small lesion in the ultrasound scanning coordinate system are obtained, and the landmark anatomical structures around the small lesion are automatically identified and marked to obtain the corresponding landmark points; the relative distance and spatial orientation relationship between the landmark points and the small lesion are recorded, and a preoperative lesion positioning reference coordinate system is established based on three or more non-collinear landmark points, and the preoperative lesion coordinates are converted to the lesion positioning reference coordinate system for storage.

[0018] The intraoperative image acquisition and registration unit is used to collect real-time conventional ultrasound images of the surgical area through the ultrasound probe carried by the robot, synchronously extract the landmark anatomical structure features in the conventional ultrasound images, and establish an intraoperative patient actual anatomical coordinate system; then, based on a multi-modal image registration algorithm, the landmark point features in the preoperative enhanced ultrasound images are matched with the landmark point features in the intraoperative conventional ultrasound images, a conversion matrix of the preoperative lesion positioning reference coordinate system and the intraoperative patient actual anatomical coordinate system is calculated, the preoperative lesion positioning reference coordinate system and the intraoperative patient actual anatomical coordinate system are aligned, and based on the aligned coordinate system, the preoperative lesion coordinates are mapped to the intraoperative image to generate a real-time ultrasound image with lesion position marking.

[0019] The lesion offset monitoring and mechanical arm adjusting unit is used for arranging infrared optical positioning sensors near the mechanical arm end and the surgical area, capturing the spatial position data of the patient body surface marker points and the mechanical arm end ablation needle in real time, calculating the lesion offset caused by the patient's breathing and body position changes by comparing the real-time collected body surface marker point position with the preoperative calibrated position, and automatically generating mechanical arm adjustment instructions to correct the ablation needle's advancing angle, depth and spatial position when the offset exceeds the preset threshold.

[0020] Further, the lesion offset monitoring and mechanical arm adjusting unit is also used for displaying the offset and adjustment trajectory in real time based on the human-computer interaction interface, and supporting manual joystick intervention adjustment.

[0021] Further, the early warning module includes a preoperative multi-modal modeling and danger calibration unit, an intraoperative vaporization area monitoring unit and a warning and linkage control unit.

[0022] The preoperative multi-modal modeling and danger calibration unit is used for synchronously collecting the ultrasound image, CT image and MRI image of the patient's tumor area by using the multi-modal image fusion technology, spatially aligning the ultrasound image, CT image and MRI image based on the gray-scale mutual information multi-modal image registration algorithm, and fusing to generate a three-dimensional integrated model; based on the three-dimensional integrated model, the three-dimensional contours of different organs are labeled by using the U-Net++ image segmentation algorithm, the corresponding dangerous areas are highlighted in different colors in the planning path output by the planning module, and the corresponding dangerous organ model is output.

[0023] The intraoperative vaporization area monitoring unit is used for collecting the intraoperative dynamic ultrasound image of the ablation area in real time by the robot-mounted ultrasonic probe, pre-processing the intraoperative dynamic ultrasound image in real time to highlight the feature difference between the vaporization area and the surrounding tissue, segmenting the pre-processed intraoperative dynamic ultrasound image frame by frame based on the deep learning-based vaporization area segmentation model, and real-time extracting the three-dimensional contour and center coordinates of the vaporization area; and based on the three-dimensional coordinates of the dangerous organs established preoperatively, the shortest straight line distance between the vaporization area edge and the dangerous organ contour is calculated in real time, a distance dynamic change curve is generated and transmitted to the planning module synchronously.

[0024] The early warning and linkage control unit is used for setting a three-level early warning mechanism based on a safety threshold.

[0025] Further, in the preoperative multi-modal modeling and danger calibration unit, differentiated safety distance thresholds are set according to the organ types; and the dangerous organ model and the tumor model are spatially superimposed and analyzed, the risk section where the ablation path intersects with the dangerous area is identified by a collision detection algorithm, and the doctor is automatically marked and prompted to adjust the path.

[0026] Further, in the early warning and linkage control unit, the three-level early warning mechanism includes: the first-level early warning only prompts non-intervention operation; the second-level early warning automatically reduces the ablation energy output power; and the third-level early warning immediately suspends the ablation operation and drives the robot to retract the ablation needle until the distance is restored to above the safety threshold.

[0027] Further, in the early warning and linkage control unit, the safety threshold is manually adjusted through the man-machine interaction interface and the non-third-level early warning linkage function is closed, and the early warning triggering time, distance data, response measures and doctor intervention operation are automatically recorded to generate a surgical safety log archive.

[0028] The above scheme has the following beneficial effects: 1. Compared with the prior art, the scheme can realize complete energy coverage of irregular tumors and effective avoidance of key anatomical structures through the precise extraction of tumor contours by the U-Net++ improved model in the planning module, the individualized simulation of the energy field driven by finite element analysis, and the optimal needle insertion scheme generated by the A* path algorithm; at the same time, the auxiliary positioning module realizes dynamic correction of small lesions with a diameter of ≤5mm through multi-modal image registration (error ≤1mm) and infrared optical tracking, solves the problem of lesion loss caused by respiration and body position changes during operation, and improves the spatial accuracy and scheme adaptability of the ablation operation as a whole;

[0029] 2. In the scheme, the early warning module constructs a three-dimensional model of dangerous organs through multi-modal image fusion and sets differentiated safety thresholds, and combines real-time monitoring of the vaporization zone driven by deep learning with graded early warning linkage control (energy regulation, operation suspension, etc.), which significantly reduces the incidence of important organ injury and other complications; at the same time, each module supports doctors to manually adjust the scheme (such as needle insertion path and early warning threshold) based on clinical experience, and automatically records the operation log, which takes into account the technical automation and clinical flexibility, and improves the standardization and operation efficiency of the surgical procedure.

[0030] Additional aspects and advantages of the application will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following and / or can be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 Flowchart of an embodiment of the ultrasound imaging guided interventional surgical robot of the application. DETAILED DESCRIPTION

[0032] The technical solutions of the present application will be described below in detail with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0033] The following detailed description illustrates the specific implementation method:

[0034] Example 1:

[0035] In current ultrasound-guided interventional procedures, the preoperative planning stage relies heavily on conventional ultrasound images, which struggle to clearly capture the boundary features of tiny lesions (≤5mm in diameter). This can lead to insufficient accuracy in recording the initial lesion location. Furthermore, during intraoperative manipulation, factors such as patient breathing and positional changes can easily cause lesion displacement, and traditional positioning methods struggle to achieve real-time registration between preoperative images and the actual intraoperative anatomical location, often resulting in lesion loss. In addition, existing interventional surgical robots lack sufficient quantification of the spatial relationship between small lesions and surrounding landmark anatomical structures, relying on the surgeon's experience to manually adjust the puncture path. This not only increases the risk of missed diagnoses but may also lead to increased damage to normal tissue due to an excessive pursuit of precise localization. Therefore, if… Figure 1 As shown, this solution specifically provides an ultrasound imaging-guided interventional surgical robot, including a planning module, an auxiliary positioning module, and an early warning module.

[0036] Specifically, the planning module includes a preoperative image acquisition and processing unit, an ablation energy field simulation unit, and a needle insertion path planning unit. The three work together to achieve preoperative assessment and plan formulation for tumor ablation.

[0037] The preoperative image acquisition and processing unit is used to perform multi-dimensional scanning of the patient's tumor area using an ultrasound probe, acquiring continuous tomographic ultrasound images. A 3D reconstruction algorithm is then used to generate a 3D ultrasound image of the tumor area. This 3D ultrasound image is further preprocessed (including grayscale correction, noise filtering, and image enhancement) to optimize the feature differences between the tumor and surrounding normal tissues, generating a first image that provides high-quality image data for subsequent contour extraction. Based on the first image, the unit uses an improved U-Net++ model for image segmentation and outputs the 3D coordinate information of the tumor, clarifying its size (volume, maximum / minimum diameter), shape (regular / irregular, clear / blurred boundaries), and spatial relationship with surrounding anatomical structures (such as blood vessels, nerves, and organs). The improved U-Net++ model uses a deep convolutional neural network to learn the grayscale distribution, texture features, and spatial structural differences between the tumor tissue and surrounding blood vessels, muscles, fat, and other tissues, automatically outlining the complete boundary of the tumor and extracting its contour. For irregularly shaped tumors, its multi-scale feature fusion structure can capture detailed structures such as the convexity and depression of the tumor edge.

[0038] The ablation energy field simulation unit is configured to construct a multi-modal ablation energy field simulation system based on the first image, call a finite element analysis algorithm, calculate the diffusion process of energy in the tumor and surrounding tissues based on a biological heat conduction equation, and generate an energy field distribution cloud map, a temperature change dynamic curve (-20-120°C), and an effective ablation range; wherein the multi-modal ablation energy field simulation system supports energy field simulation of mainstream ablation modes such as radiofrequency ablation, microwave ablation, and cryoablation; and the doctor can input or adjust the ablation parameters through the human-computer interaction interface, including the ablation energy intensity (0-300 W), the action time (10-300 s), the number of ablation needles (1-4), and the needle path distribution mode.

[0039] The needle insertion path planning unit is configured to combine the tumor contour data, the energy field simulation result, and the important anatomical structures labeled preoperatively, and to automatically generate an optimal needle insertion path based on an optimization basic strategy using an A * The path search algorithm automatically generates an optimal needle insertion path, and outputs the needle insertion point coordinates, the puncture angle, and the push depth of the ablation needle. In the planning process, the optimization objectives are that the energy field completely covers the tumor, the path avoids dangerous structures, and the puncture depth is the shortest. In addition, the system supports the doctor to manually adjust the path based on clinical experience through dragging, rotating, and other methods, and after adjustment, the system updates the energy field simulation result and the risk assessment report (such as normal tissue damage probability) in real time to assist the doctor to confirm the final scheme.

[0040] Specifically, the assisted positioning module includes a preoperative contrast imaging and coordinate system establishment unit, an intraoperative image acquisition and registration unit, and a lesion shift monitoring and mechanical arm adjustment unit, and realizes accurate positioning of small lesions through a closed-loop process of “preoperative calibration-intraoperative tracking-dynamic correction”;

[0041] The preoperative contrast imaging and coordinate system establishment unit is configured to preoperatively inject an ultrasound contrast agent intravenously, perform three-dimensional scanning on the lesion area at a frame rate of 5-30 Hz and a spatial resolution of 0.1-0.3 mm after the contrast agent is enriched in the lesion area (1-5 minutes after injection, adjusted according to the blood supply characteristics of the lesion), collect enhanced ultrasound images of the lesion and surrounding tissues, and highlight the gray difference between the lesion and normal tissues with a diameter of ≤5 mm; pre-process the enhanced ultrasound images (including adaptive noise reduction, contrast enhancement, and artifact removal), extract the three-dimensional contour of the small lesion through a U-Net++ improved model, obtain the initial position coordinates (X0, Y0, Z0) of the small lesion in the ultrasound scanning coordinate system, automatically identify and mark the landmark anatomical structures (such as vertebral bodies and bifurcations of large blood vessels) around the small lesion, obtain the corresponding landmark points, record the relative distance and spatial orientation relationship between the landmark points and the small lesion, and establish a preoperative lesion positioning reference coordinate system based on three or more non-collinear landmark points, and store the preoperative lesion coordinates in the lesion positioning reference coordinate system;

[0042] The intraoperative image acquisition and registration unit is used for real-time acquisition of a conventional ultrasound image of a surgical region by an ultrasonic probe carried by a robot, synchronous extraction of a landmark anatomical structure feature in the conventional ultrasound image, and establishment of an actual anatomical coordinate system of a patient during surgery; based on a multi-modal image registration algorithm, a landmark point feature in a preoperative enhanced ultrasound image is matched with a landmark point feature of the intraoperative conventional ultrasound image, a conversion matrix of the preoperative lesion positioning reference coordinate system and the actual anatomical coordinate system of the patient during surgery is calculated, alignment of the preoperative lesion positioning reference coordinate system and the actual anatomical coordinate system of the patient during surgery is performed, based on the aligned coordinate system, a preoperative lesion coordinate is mapped to an intraoperative image, and a real-time ultrasound image with a lesion position marker is generated for a doctor to intuitively view the lesion position;

[0043] The lesion shift monitoring and mechanical arm adjustment unit is used for arranging an infrared optical positioning sensor near the end of the mechanical arm and the surgical region, real-time capture of spatial position data of a patient body surface marker (having a fixed anatomical relationship with the lesion) and an ablation needle at the end of the mechanical arm, calculation of a lesion shift amount (ΔX, ΔY, ΔZ) caused by changes in the patient's breathing and body position by comparing the real-time acquired body surface marker position with a preoperative calibrated position, automatic generation of a mechanical arm adjustment instruction when the shift amount exceeds a preset threshold (0.5-1 mm, which can be adjusted according to the size of the lesion), so as to control the robot to correct the advancing angle, depth and spatial position of the ablation needle, and ensure that the ablation needle is always aligned with the center of the lesion; at the same time, the human-computer interaction interface displays the shift amount and the adjustment trajectory in real time, and supports the doctor to intervene in adjustment through a manual joystick.

[0044] Specifically, the early warning module includes a preoperative multi-modal modeling and danger calibration unit, an intraoperative vaporization zone monitoring unit, and an early warning and linkage control unit.

[0045] The preoperative multi-modal modeling and danger calibration unit is used for synchronous acquisition of ultrasound images, CT images and MRI images of a tumor region of a patient by using a multi-modal image fusion technology, spatial alignment of the ultrasound images, the CT images and the MRI images based on a multi-modal image registration algorithm based on gray-scale mutual information, and fusion to generate a three-dimensional comprehensive model; based on the three-dimensional comprehensive model, a U-Net++ image segmentation algorithm is used to respectively segment three-dimensional contours of different organs (for example, important organs such as an intestinal canal, a trachea, a portal vein and a aorta), different color highlighting markers are marked in a corresponding dangerous region in a planning path output by a planning module, and a corresponding dangerous organ model is output; at the same time, a difference safety distance threshold is set according to the types of the organs (for example, the safety threshold is set to 0.5-1 mm for hollow organs such as the intestinal canal and the trachea, and the safety threshold is set to 0.2-0.3 mm for blood-rich organs such as large blood vessels), the dangerous organ model is spatially superimposed with a tumor model, a risk section where an ablation path intersects with a dangerous region is identified by a collision detection algorithm, and a doctor is automatically prompted to adjust the path.

[0046] The intraoperative vaporization area monitoring unit is used for real-time acquisition of intraoperative dynamic ultrasound images of the ablation area by a robot-mounted ultrasonic probe, real-time preprocessing of the intraoperative dynamic ultrasound images, highlighting of feature differences between the vaporization area and surrounding tissues, frame-by-frame segmentation of the preprocessed intraoperative dynamic ultrasound images by using a vaporization area segmentation model based on deep learning, real-time extraction of three-dimensional contours and center coordinates of the vaporization area, real-time calculation of the shortest straight line distances between the vaporization area edge and the contours of each dangerous organ based on the three-dimensional coordinates of the dangerous organs established before the operation by using the Euclidean distance algorithm, generation of distance dynamic change curves and synchronous transmission to the planning module.

[0047] The early warning and linkage control unit is used for setting a three-level early warning mechanism based on a safety threshold, allowing doctors to manually adjust the safety threshold and close the non-three-level early warning linkage function through a human-computer interaction interface, and automatically recording early warning trigger time, distance data, response measures and doctor intervention operations, and generating a surgical safety log archive. The three-level early warning mechanism includes: first-level warning only prompts non-intervention operation; second-level warning automatically reduces ablation energy output power; third-level warning immediately suspends ablation operation and drives the robot to retract the ablation needle until the distance is restored to above the safety threshold. For the setting of the safety threshold, for example: when the distance between the vaporization area and the dangerous organ is in the range of "safety threshold-safety threshold x 1.2", first-level warning is triggered (green indicator light flickering + low-frequency prompt sound); when the distance is in the range of "safety threshold x 0.8-safety threshold", second-level warning is triggered (yellow indicator light always on + medium-frequency prompt sound + interface pop-up window prompt); when the distance < safety threshold x 0.8, third-level warning is triggered (red indicator light flashing + high-frequency alarm sound + forced pop-up window locking).

[0048] Embodiment 2:

[0049] The difference from the above embodiments is that, as Figure 1As shown, on the basis of retaining the original three-level warning mechanism (level one prompt, level two energy adjustment, and level three pause and retreat) of the early warning and linkage control unit and the dynamic monitoring function of the dangerous area, the early warning and linkage control unit is also used to calculate the credibility (expressed in percentage) of complete tumor inactivation after each ablation by a deep learning model (a training model integrating tumor biological characteristics, ablation energy parameters, and image features) through the preoperative planned ablation range parameters (such as tumor volume and preset ablation boundary) and the real-time monitored vaporization coverage range, combined with the post-ablation contrast-enhanced ultrasound image features (such as the proportion of non-enhanced area and contrast agent inflow speed), to provide a quantitative reference for postoperative surgical plan adjustment, thereby effectively solving the problems in traditional ablation, such as “discovering incomplete inactivation after surgery and needing secondary surgery”, “relying on the subjective experience of doctors for supplementary operation”, and “lack of individualized evaluation criteria for different tumors”, improving the complete inactivation rate of tumors in clinical application, reducing the risk of short-term recurrence after surgery, and effectively reducing the range of normal tissue damage, thereby effectively shortening the treatment cycle of patients, improving the treatment effect, and taking into account the safety and efficiency of surgery. In addition, through standardized operation processes driven by data, the homogeneity level of complex tumor ablation surgery in primary medical institutions can also be effectively improved.

[0050] Specifically, when the three-level warning is triggered for the first time, the ablation time reaches the preset value (such as 60 s), or the vaporization area covers the preset range (such as 120% of the tumor volume), the postoperative contrast-enhanced ultrasound image of the ablation area is automatically collected, the spatial overlap of the non-enhanced area (inactivated area) and the original tumor area is extracted through image segmentation, combined with the preoperative tumor blood supply characteristics, the cumulative value of ablation energy, and the record of intraoperative offset adjustment, the credibility of complete tumor inactivation of this ablation (70%-100%) is comprehensively calculated.

[0051] If the evaluation result shows that the credibility is less than 85%-95% (this threshold value can be individually set according to the pathological type of tumor, such as 85% for hepatocellular carcinoma and 90% for cholangiocellular carcinoma), it is determined that the patient has a high risk of tumor recurrence after surgery, and the “suggestion of supplementary ablation” is prompted in the pop-up window of the human-computer interaction interface. Then, based on the position of the remaining survival area (enhanced area), combined with the safety threshold of dangerous organs, the recommended path of supplementary ablation (such as adjusting the needle insertion angle by 10° and increasing the energy to 90 W) is generated.

[0052] After the doctor performs the second ablation based on the recommended path of supplementary ablation, the contrast-enhanced ultrasound image is collected again, the vaporization area range and energy parameters of the two ablations are superimposed, and the credibility of complete tumor inactivation is re-evaluated until the credibility is greater than 90%.

[0053] Obviously, the above embodiments are merely example for clearly illustrating but not limitation to the embodiments. Based on the above description, other different forms of changes or variations can be made by those skilled in the art. Here, all the embodiments need not and can not be enumerated. The obvious changes or variations derived from this still fall within the protection scope of the present application.

Claims

1. An ultrasound imaging-guided interventional surgical robot, characterized in that, include: The planning module is used to extract tumor contour features and determine tumor feature information based on three-dimensional ultrasound images of the patient's tumor area, and to construct an ablation energy field simulation model. It calculates the energy field simulation results under different ablation parameters and automatically generates the optimal needle insertion path planning scheme based on the energy field simulation results. The auxiliary positioning module is used to obtain the initial three-dimensional spatial coordinates of the lesion based on the enhanced image of the lesion, and at the same time record the relative positional relationship between the lesion and the surrounding landmark anatomical structures to establish a lesion positioning reference coordinate system; it is also used to obtain the patient's actual anatomical coordinate system during the operation, and align the lesion positioning reference coordinate system established before the operation with the patient's actual anatomical coordinate system during the operation to perform real-time dynamic tracking of the lesion position. The early warning module is used to establish a 3D model of normal organs surrounding the tumor, import the 3D model of normal organs into the ablation energy field simulation model, mark dangerous areas on the 3D model of normal organs, set corresponding safe distance thresholds, analyze the potential intersection between the ablation movement path generated by the ablation energy field simulation model and the dangerous areas, and optimize the ablation movement path. It is also used to monitor changes in the vaporization zone of the ablation area in real time during the operation, extract the outline and location information of the vaporization zone based on the vaporization zone segmentation model, calculate the real-time distance between the vaporization zone and the pre-marked dangerous area, compare it with the safe distance threshold, and trigger the alarm strategy.

2. The ultrasound imaging-guided interventional surgical robot according to claim 1, characterized in that: The planning module includes a preoperative image acquisition and processing unit, an ablation energy field simulation unit, and a needle insertion path planning unit; The preoperative image acquisition and processing unit is used to perform multi-dimensional scanning of the patient's tumor area using an ultrasound probe, acquire continuous high frame rate tomographic ultrasound images, generate a three-dimensional ultrasound image of the tumor area using a three-dimensional reconstruction algorithm, and then preprocess the three-dimensional ultrasound image of the tumor area to optimize the feature differences between the tumor and the surrounding normal tissue to generate a first image; it is also used to perform image segmentation based on the first image using an improved U-Net++ model, and output the three-dimensional coordinate information of the tumor to clarify the size of the tumor and its spatial positional relationship with the anatomical structures of surrounding organs; The ablation energy field simulation unit is used to construct a multimodal ablation energy field simulation system based on the first image, call the finite element analysis algorithm, calculate the diffusion process of energy in the tumor and surrounding tissues based on the biological heat conduction equation, and generate energy field distribution cloud map, temperature change dynamic curve and effective ablation range; The needle insertion path planning unit combines tumor contour data, energy field simulation results, and preoperatively annotated important anatomical structures, and, based on an optimization strategy, employs A... * The path search algorithm automatically generates the optimal needle insertion path and outputs the needle insertion point coordinates, puncture angle, and advancement depth.

3. The ultrasound imaging-guided interventional surgical robot according to claim 2, characterized in that: In the needle insertion path planning unit, the optimization basic strategies include: the energy field completely covers the tumor, the path avoids dangerous structures, and the shortest puncture depth.

4. The ultrasound imaging-guided interventional surgical robot according to claim 3, characterized in that: The needle insertion path planning unit is also used to support manual adjustment of the optimal needle insertion path planning scheme. After adjustment, the energy field simulation results and risk assessment reports are updated in real time to help confirm the final planning scheme.

5. The ultrasound imaging-guided interventional surgical robot according to claim 4, characterized in that: The auxiliary positioning module includes a preoperative angiography and coordinate system establishment unit, an intraoperative image acquisition and registration unit, and a lesion displacement monitoring and robotic arm adjustment unit; The preoperative contrast imaging and coordinate system establishment unit is used to perform three-dimensional scanning of the lesion area by intravenously injecting ultrasound contrast agent before surgery. After the contrast agent accumulates in the lesion area, the ultrasound probe is controlled to acquire enhanced ultrasound images of the lesion and surrounding tissues, highlighting the grayscale difference between lesions with a diameter ≤5mm and normal tissues. The enhanced ultrasound images are preprocessed, and the three-dimensional contour of small lesions is extracted using the U-Net++ improved model. The initial position coordinates of the small lesions in the ultrasound scanning coordinate system are obtained. Then, the landmark anatomical structures around the small lesions are automatically identified and marked to obtain the corresponding landmark points. The relative distance and spatial orientation relationship between the landmark points and the small lesions are recorded. Based on three or more non-collinear landmark points, a preoperative lesion positioning reference coordinate system is established, and the preoperative lesion coordinates are converted to the lesion positioning reference coordinate system and stored. The intraoperative image acquisition and registration unit is used to acquire conventional ultrasound images of the surgical area in real time using an ultrasound probe mounted on the robot, simultaneously extract landmark anatomical features from the conventional ultrasound images, and establish the patient's actual anatomical coordinate system during the operation. Then, based on a multimodal image registration algorithm, the landmark features in the preoperative enhanced ultrasound images are matched with the landmark features in the intraoperative conventional ultrasound images. The transformation matrix between the preoperative lesion localization reference coordinate system and the patient's actual anatomical coordinate system during the operation is calculated, and the preoperative lesion localization reference coordinate system is aligned with the patient's actual anatomical coordinate system during the operation. Based on the aligned coordinate system, the preoperative lesion coordinates are mapped to the intraoperative images to generate real-time ultrasound images with lesion location markers. The lesion deviation monitoring and robotic arm adjustment unit is used to deploy infrared optical positioning sensors at the end of the robotic arm and near the surgical area to capture the spatial position data of the patient's surface markers and the ablation needle at the end of the robotic arm in real time. By comparing the real-time collected surface marker positions with the pre-operative calibration positions, the lesion deviation caused by changes in the patient's breathing and body position is calculated. When the deviation exceeds a preset threshold, the robotic arm adjustment command is automatically generated to correct the advancement angle, depth and spatial position of the ablation needle.

6. The ultrasound imaging-guided interventional surgical robot according to claim 5, characterized in that: The lesion offset monitoring and robotic arm adjustment unit is also used to display the offset and adjustment trajectory in real time based on the human-machine interface, and supports adjustment by manual crank.

7. The ultrasound imaging-guided interventional surgical robot according to claim 6, characterized in that: The early warning module includes a preoperative multimodal modeling and hazard calibration unit, an intraoperative vaporization zone monitoring unit, and an early warning and linkage control unit; The preoperative multimodal modeling and risk assessment unit is used to simultaneously acquire ultrasound, CT and MRI images of the patient's tumor area using multimodal image fusion technology. Based on the gray-scale mutual information multimodal image registration algorithm, the ultrasound, CT and MRI images are spatially aligned and fused to generate a three-dimensional comprehensive model. Based on the three-dimensional integrated model, the three-dimensional contours of different organs are marked by the U-Net++ image segmentation algorithm. The corresponding dangerous areas are highlighted with different colors in the planning path output by the planning module, and the corresponding dangerous organ models are output. The intraoperative vaporization zone monitoring unit is used to acquire real-time dynamic ultrasound images of the ablation area via an ultrasound probe mounted on the robot. Real-time preprocessing of these images highlights the differences between the vaporization zone and surrounding tissues. A deep learning-based vaporization zone segmentation model is used to segment the preprocessed dynamic ultrasound images frame by frame, extracting the 3D contour and center coordinates of the vaporization zone in real time. Based on the preoperatively established 3D coordinates of the critical organs, the unit calculates the shortest straight-line distance between the edge of the vaporization zone and the contours of each critical organ in real time, generating a dynamic distance change curve and transmitting it synchronously to the planning module. The early warning and linkage control unit is used to set up a three-level early warning mechanism based on safety thresholds.

8. The ultrasound imaging-guided interventional surgical robot according to claim 7, characterized in that: In the preoperative multimodal modeling and risk assessment unit, differentiated safety distance thresholds are set according to organ type; and the dangerous organ model and the tumor model are spatially overlaid for analysis. The collision detection algorithm identifies the risk segments where the ablation path intersects with the dangerous area, automatically marks them, and prompts the doctor to adjust the path.

9. The ultrasound imaging-guided interventional surgical robot according to claim 8, characterized in that: In the early warning and linkage control unit, the three-level early warning mechanism includes: Level 1 early warning only prompts the operator to not intervene; Level 2 early warning automatically reduces the ablation energy output power; Level 3 early warning immediately suspends the ablation operation and drives the robot to retract the ablation needle until the distance is restored to above the safe threshold.

10. The ultrasound imaging-guided interventional surgical robot according to claim 9, characterized in that: In the early warning and linkage control unit, the safety threshold can be manually adjusted and the non-level 3 early warning linkage function can be turned off through the human-computer interaction interface. The system automatically records the early warning trigger time, distance data, response measures and doctor intervention operations, and generates a surgical safety log archive.

Images