Orthopedic surgery robot intelligent guiding system and method based on mechanical attribute map

Abstract

The invention discloses an intelligent guiding system and method for an orthopedic surgery robot based on a mechanical attribute map, and the system comprises an image processing module which is used for calculating the standardized Henry unit value of each voxel according to the preoperative three-dimensional medical image data of a patient; the mechanical attribute conversion module is used for converting the standardized Hensword unit value into a skeleton mechanical attribute value based on a Hensword unit-mechanical attribute calibration model, and generating a mechanical attribute map; the map visualization module is used for displaying the mechanical attribute map in a color mapping mode and displaying the mechanical attribute map and a skeleton three-dimensional model in an overlapping mode to obtain a fusion visualization map; and the path planning and analyzing module is used for automatically planning a bone entering path based on the mechanical attribute map and the operation requirement or interactively planning a bone entering path based on the fusion visual map. The system can solve the problem that an existing orthopedic robot navigation system only has geometric navigation and lacks mechanical predictability and safety guidance.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, specifically to an intelligent guidance system and method for orthopedic surgical robots based on mechanical property maps, and in particular to a surgical navigation and control system that integrates medical image processing, biomechanical modeling and real-time force sensing. Background Technology

[0002] Orthopedic surgical robots, as an important branch of medical robotics technology, have been widely used in recent years in fields such as spinal surgery and joint replacement due to their ability to improve surgical precision, reduce human error, and enable minimally invasive procedures. These systems typically integrate core modules such as preoperative planning, navigation and positioning, and surgical instrument control to assist operators in precisely performing surgical procedures. Especially in spinal surgery, procedures such as pedicle screw placement highly depend on the precise preparation of the surgical channel for success.

[0003] Currently, the technology behind orthopedic surgical robots primarily relies on geometric path planning and navigation based on medical images. Typically, a 3D reconstruction is performed preoperatively using the patient's CT or CBCT images. The operator plans the ideal bone-entry path on this 3D model. This path defines the geometric trajectory and spatial boundaries of the instrument's entry into the bone, ensuring that the path avoids important nerves and blood vessels and reaches the target area. During the surgery, certain points in the instrument's bone-entry process (e.g., encountering high resistance) require medical imaging scans (usually X-ray fluoroscopy or intraoperative CT) to verify the accuracy of the actual bone-entry path.

[0004] In existing technologies, some research has begun to focus on mechanical signals during surgery. For example, some approaches attempt to monitor cutting forces during bone drilling using a six-dimensional force / torque sensor integrated into the end of the surgical instrument, and to adjust the feed rate using control algorithms (such as PID controllers) to maintain a constant drilling force, aiming to reduce thermal damage to bone tissue and improve safety. Other patent documents disclose methods for calculating external forces acting on surgical instruments by using markings (such as optical target balls) mounted on the instruments and combining this with a mechanical model of the instruments.

[0005] Although existing technologies have achieved precise navigation based on geometric space, they still have significant limitations when dealing with the complex biomechanical environment of real surgery. These limitations mainly lie in the inability to predict individual differences in bone biomechanical properties and the lack of deep integration between biomechanical feedback and geometric navigation.

[0006] Firstly, existing navigation systems based on geometric path planning primarily rely on skeletal imaging data, such as preoperative CT scans, for planning. This approach fails to reflect the individualized differences in the mechanical properties of bones among different patients, and even among different parts of the same patient. Bone density and hardness are related to their imaging Henle values, and these properties exhibit significant individual and regional variations. The system cannot inform the operator preoperatively of potentially abnormally hard bone areas along the planned path, forcing the operator to react passively only when unexpected high resistance is encountered during surgery, relying on tactile feedback. This lack of foresight significantly increases surgical risks.

[0007] Secondly, while some navigation systems can monitor cutting forces and other mechanical signals in real time, this mechanical information is not effectively correlated and integrated with preoperative imaging information, constituting passive monitoring. The navigation interface mainly provides geometric position information of surgical instruments, but lacks interpretation and early warning of the current skeletal biomechanical state during operation. Therefore, when a sudden increase in drill force is detected, the system can only respond passively (such as slowing down or stopping), unable to determine from the preoperative data whether the high resistance is due to "an expected high-hardness bone area" or "an unexpected path deviation or instrument failure." This lack of discriminatory ability limits the intelligence level of the guidance system and also affects the smoothness of the surgery due to the need to interrupt bone entry checks.

[0008] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0009] Purpose of the invention: The technical problem to be solved by the present invention is to provide an intelligent guidance system and method for orthopedic surgical robots based on mechanical property maps, which addresses the shortcomings of existing orthopedic robots that only rely on geometric navigation and lack mechanical predictability and safe guidance.

[0010] To address the aforementioned technical problems, this invention discloses an intelligent guidance system for orthopedic surgical robots based on a biomechanical property map. The system includes:

[0011] The image processing module is used to calculate the standardized Henle unit values ​​of each voxel based on the patient's preoperative three-dimensional medical image data.

[0012] The mechanical property conversion module is used to convert the standardized Henry's unit values ​​into skeletal mechanical property values ​​based on the "Henry's unit-mechanical property" calibration model, and generate a mechanical property map.

[0013] The map visualization module is used to display the mechanical property map in a color mapping manner and overlay it with the skeletal 3D model to obtain a fused visualization map.

[0014] And a path planning and analysis module, used to automatically or interactively plan the bone entry path based on the biomechanical property map and surgical requirements, or based on the fused visualization map.

[0015] The system disclosed in this invention expands the navigation dimension of surgical robots from a single geometric space to a mechanical property space, enabling preoperative prediction and visualization of bone mechanical properties by constructing a mechanical property map. This improvement addresses the core deficiency of existing surgical robot navigation systems that rely solely on morphological information and cannot predict individual differences in bone mechanical properties. It intuitively presents bone mechanical performance information that traditional imaging (such as X-rays) cannot provide by integrating a visualization map; it transforms complex and abstract mechanical concepts into intuitive and operable visual information, greatly reducing the cognitive burden on operators and improving the efficiency and accuracy of research and decision-making. Through a path planning and analysis module, it plans the bone entry path based on the aforementioned mechanical property map, elevating bone entry path planning from "avoiding geometrically dangerous structures" to "anticipating mechanically difficult areas," realizing a shift from passive mechanical response to active planning. High-hardness areas on the planned path can be identified preoperatively, allowing for targeted strategies (such as pre-selecting a better path or adjusting operating parameters), improving the predictability and safety of the surgery, and reducing the risks of instrument deviation, tissue damage, and drill jamming due to unexpected encounters with hard bone.

[0016] In one embodiment, the mechanical property value is bone stiffness or elastic modulus.

[0017] In one embodiment, the path planning and analysis module is further configured to synchronously generate a predicted cutting force curve for the path based on the mechanical property values ​​of the area traversed by the bone entry path in the mechanical property map, and to provide the predicted cutting force curve to the operator in a visual manner.

[0018] In one embodiment, the system further includes:

[0019] The real-time monitoring and comparison module is used to acquire real-time cutting force data of the surgical instrument during the feeding process along the bone entry path and compare it with the data at the same position of the predicted cutting force curve in real time.

[0020] And an early warning and / or control module for performing at least one intraoperative guidance operation; the intraoperative guidance operation includes, but is not limited to: providing the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve formed by the real-time cutting force data and the predicted cutting force curve; outputting early warning information based on the comparison results; and adaptively adjusting the feed speed and / or rotation speed of the surgical instrument through the surgical instrument power system in combination with the current position mechanical property value and the comparison results.

[0021] In one embodiment, the surgical requirements are input by the operator and are selected from bone tissue sampling, pedicle screw placement, or vertebral body distraction.

[0022] In one embodiment, the step of outputting warning information based on the comparison result includes:

[0023] When the real-time cutting force data continues for a specified duration for a number of times exceeding the corresponding predicted value and reaching different multiple thresholds, an alarm of the corresponding level is triggered;

[0024] The duration of the specified duration is determined based on the data transmission frequency, and the duration of the specified duration is set to be greater than the transmission period of a single data point.

[0025] This embodiment addresses the problem of false alarms triggered by instantaneous deviations by filtering based on duration and multiple thresholds, thereby improving the stability of early warnings.

[0026] In one embodiment, the different levels of alarms include:

[0027] Level 1 Alarm: When the real-time cutting force data continuously exceeds the first preset multiple of the predicted value but does not exceed the second preset multiple of the predicted value, a yellow visual warning is triggered, and a warning icon is displayed on the terminal interface;

[0028] Level 2 Alarm: When the real-time cutting force data continuously exceeds the second preset multiple of the predicted value, a red visual warning is triggered, accompanied by an audible alarm, and a pause command is sent to the surgical instrument power system.

[0029] This embodiment avoids abrupt shutdowns through tiered processing: when the real-time cutting force data continuously exceeds the first preset multiple of the predicted value but does not exceed the second preset multiple of the predicted value, the system is considered low-risk and only requires a warning. If the operator subjectively judges that the risk is high, they can pause the operation themselves (by pressing the emergency stop button, releasing the foot pedal switch, etc.); when the real-time cutting force data continuously exceeds the second preset multiple of the predicted value for a specified period of time, it may be due to path deviation, instrument failure, or other reasons, and continuing the operation is high-risk. In this case, the surgical instrument inspection is immediately suspended, balancing safety and efficiency.

[0030] In one embodiment, the early warning and / or control module adaptively adjusts the feed speed and / or rotation speed of the surgical instrument through the surgical instrument power system according to a preset hierarchical control rule;

[0031] The preset hierarchical control rules include:

[0032] When it is predicted that the surgical instrument will enter a high-hardness area, the warning and / or control module reduces the feed speed of the surgical instrument and increases the rotation speed of the surgical instrument through the surgical instrument power system, so that the operation process is smooth and controllable, while protecting the bone and the surgical instrument.

[0033] When the real-time cutting force data continues for a specified duration exceeding a preset multiple of the predicted value, the warning and / or control module further reduces the feed speed of the surgical instrument through the surgical instrument power system;

[0034] When a low-hardness region is predicted to be entered, the warning and / or control module increases the feed speed of the surgical instrument through the surgical instrument power system to improve surgical efficiency.

[0035] In one embodiment, the map visualization module is further configured to provide a transparency adjustment function and a point-and-click query function for adjusting the overlay ratio of the mechanical property map and the skeletal 3D model display.

[0036] In one embodiment, the interactive planning of the bone entry path based on the fused visualization map specifically involves: after the operator manually plans the bone entry route on the fused visualization map, the path planning and analysis module automatically analyzes the mechanical property values ​​of the areas traversed by the bone entry path and generates a mechanical risk assessment report; the mechanical risk assessment report includes at least the total path length, the expected average cutting force, the location and value of the peak cutting force, and the proportion of high-hardness areas;

[0037] And / or, the automatic planning of bone entry paths based on the biomechanical property map and surgical requirements specifically involves the path planning and analysis module generating multiple recommended bone entry paths with different focuses on safety, biomechanical load, and surgical efficiency by minimizing the comprehensive cost function based on the biomechanical property map and surgical requirements, and outputting them for the operator to select.

[0038] In one embodiment, the system integrates a force sensor and a navigation and positioning device. The force sensor is used to collect real-time cutting force data of the surgical instrument. The navigation and positioning device is used to simultaneously calculate the real-time spatial pose of the surgical instrument. The force sensor and the navigation and positioning device are respectively communicatively connected to the real-time monitoring and comparison module, enabling the real-time monitoring and comparison module to obtain the real-time cutting force data and real-time spatial pose of the surgical instrument from the force sensor and the navigation and positioning device, respectively.

[0039] In one embodiment, the system further includes a data archiving and model iteration module. This module, while adhering to ethical and privacy regulations, anonymizes and stores equipment information, dosage settings, mechanical property prediction data, and intraoperative mechanical feedback data for each surgery. It also periodically retrains and optimizes the "Henness unit-mechanical property" calibration model using the accumulated data. Intraoperative mechanical feedback data includes, but is not limited to, real-time cutting force data of surgical instruments. Mechanical property prediction data includes, but is not limited to, the predicted cutting force curve of the bone entry path.

[0040] A second aspect of the present invention provides an intelligent guidance method for orthopedic surgical robots based on a mechanical property map, the method comprising:

[0041] The standardized Henlein unit values ​​for each voxel were calculated based on the patient's preoperative three-dimensional medical imaging data.

[0042] Based on the "Henle unit-mechanical property" calibration model, standardized Henle unit values ​​are converted into skeletal mechanical property values ​​to generate a mechanical property map;

[0043] The mechanical property map is displayed using a color mapping method and overlaid with the 3D skeletal model to obtain a fused visual map.

[0044] Based on the mechanical property map and surgical requirements, the bone entry path is automatically or interactively planned based on the fused visualization map, and a predicted cutting force curve is generated by combining the mechanical property values ​​of the tissue areas traversed by the bone entry path in the mechanical property map.

[0045] The real-time cutting force data of the surgical instrument during the feeding process along the bone entry path is obtained and compared in real time with the data at the same position of the predicted cutting force curve;

[0046] Perform at least one of the intraoperative guidance operations; the intraoperative guidance operations include, but are not limited to: providing the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve formed by the real-time cutting force data and the predicted cutting force curve; outputting early warning information based on the comparison results; and adaptively adjusting the feed speed and / or rotation speed of the surgical instrument through the surgical instrument power system in combination with the current position mechanical attribute value and the comparison results, so as to realize the closed-loop linkage guidance of preoperative mechanical prediction and intraoperative mechanical feedback.

[0047] Beneficial effects:

[0048] 1. This invention obtains a biomechanical property map based on preoperative three-dimensional medical images; by visualizing the biomechanical property map with color levels and overlaying it with a bone model, a fused visualization map is obtained; the fused visualization map intuitively presents bone biomechanical performance information that traditional images (such as X-rays) cannot provide; it transforms complex and abstract mechanical concepts into intuitive and operable visual information, greatly reducing the cognitive burden on operators and improving the efficiency and accuracy of research and decision-making.

[0049] 2. By using a biomechanical property map, the distribution of bone stiffness along the bone entry path was predicted, solving the problem that existing navigation systems cannot predict bone biomechanical properties, resulting in a lack of biomechanical foresight during surgery.

[0050] 3. The path planning and analysis module plans the bone entry path based on the aforementioned biomechanical property map, elevating bone entry path planning from "avoiding geometrically dangerous structures" to "anticipating biomechanically challenging areas," thus shifting from a passive biomechanical response to proactive planning. More specifically, in the system's automatic planning mode, the path planning and analysis module, based on the biomechanical property map and surgical requirements, generates multiple recommended bone entry paths with different emphases on safety, biomechanical load, and surgical efficiency by minimizing the comprehensive cost function, and outputs these paths for the operator to select.

[0051] 4. The path planning and analysis module synchronously generates a predicted cutting force curve for the area traversed by the bone entry path in the mechanical property map based on the mechanical property values ​​of the area. The predicted cutting force curve is then visualized and provided to the operator, making the invisible internal mechanical environment of the bone predictable. This allows the operator to have a preliminary understanding and psychological expectation of the mechanical situation they will face before the surgery begins.

[0052] 5. By using the predicted cutting force curve as a reference baseline, the early warning and / or control module can overlay and present the real-time cutting force data of the surgical instruments on the terminal interface, upgrading the surgery from a "qualitative" operation that relies on the operator's personal experience and vague senses to a "quantitative and visualized" precision operation that is guided by objective data and provides real-time feedback.

[0053] 6. By using an early warning and / or control module to classify and process the degree of real-time cutting force deviation, the system avoids "one-size-fits-all" shutdowns. Low-risk cases only require an alert, which can avoid unnecessary interruptions caused by a brief, acceptable increase in force (such as penetrating a known cortical bone) (as confirmed by traditional intraoperative scanning), improving the continuity of operation and reducing the total operation time extension caused by false alarms. High-risk cases require a forced pause to avoid the risks of path deviation or instrument failure, balancing safety and efficiency.

[0054] Meanwhile, the early warning and / or control module provides graded alarms based on the degree of real-time chip force deviation, helping operators distinguish between "expected high resistance" and "unexpected path deviation or machine failure," providing objective mechanical guidance standards for younger operators, making it easier for them to understand and make decisions.

[0055] 7. Compared to simply outputting warning information, the warning and / or control module adaptively adjusts the feed speed and / or rotation speed of the surgical instrument based on the mechanical property values ​​of the current area in the mechanical property map and the deviation between the real-time cutting force data and the predicted values. This significantly improves the smoothness and stability of the surgery, balancing safety and surgical efficiency.

[0056] 8. By setting up data archiving and model iteration modules, a closed-loop learning system of "data acquisition - model optimization - performance improvement" was constructed. This allows the "Henness unit - mechanical property" calibration model to continuously self-optimize with the accumulation of clinical applications, constantly enhancing its predictive accuracy. This improvement can adapt to problems in different populations, equipment, and pathological conditions through long-term accumulation, possessing the ability for continuous system evolution and long-term clinical applicability. Unlike traditional one-time calibration systems, this system can continuously accumulate and analyze actual clinical data, making the predictive model increasingly accurate and capable of personalized predictions for individual differences, ensuring the long-term leading performance and added value of the system. Attached Figure Description

[0057] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0058] Figure 1 This is a system block diagram of an intelligent guidance system for orthopedic surgical robots based on a mechanical property map, provided in the first embodiment of the present invention.

[0059] The accompanying figure labels are explained as follows:

[0060] 000. Surgical instruments; 110. Image processing module; 120. Mechanical property conversion module; 130. Path planning and analysis module; 140. Real-time monitoring and comparison module; 150. Early warning and / or control module; 160. Map visualization module; 170. Data archiving and model iteration module; 180. Force sensor; 190. Navigation and positioning device; 200. Surgical instrument power system. Detailed Implementation

[0061] The applicant found that existing orthopedic surgical robots are mainly based on image-based geometric navigation, which lacks the ability to predict bone changes and provide intelligent safety guidance during surgery. They can only respond passively, leading to increased operational risks and a less smooth workflow.

[0062] According to the applicant's analysis, the reason for the above problems is that the existing system cannot predict individual differences in skeletal biomechanics, and the intraoperative biomechanical feedback is not integrated with the preoperative geometric navigation.

[0063] Therefore, this invention provides an intelligent guidance system for orthopedic surgical robots based on a mechanical property map. The system obtains a mechanical property map based on preoperative three-dimensional medical images; displays the mechanical property map in a color mapping manner and overlays it with a three-dimensional bone model to obtain a fused visualization map; and automatically or interactively plans the bone entry path based on the mechanical property map and surgical requirements, thereby upgrading the surgical planning criteria from a single morphological safety to a comprehensive intelligent decision that integrates mechanical safety and mechanical efficiency.

[0064] Furthermore, by predicting the cutting force based on the mechanical property map and bone entry path, and comparing it with the real-time cutting force during surgery to form an early warning and / or adaptive control closed loop for surgical instruments, the problem of existing orthopedic robots relying solely on geometric navigation and lacking mechanical predictability and safety guidance can be solved.

[0065] The system of this invention achieves a leap from "simple geometric path planning" to "active map navigation with geometric and mechanical properties".

[0066] The present invention will now be described in detail with reference to the embodiments.

[0067] Example 1

[0068] like Figure 1 As shown, an embodiment of the present invention provides an intelligent guidance system for orthopedic surgical robots based on a biomechanical property map. The system includes an image processing module 110, a biomechanical property conversion module 120, a map visualization module 160, and a path planning and analysis module 130. Specifically, the image processing module 110 calculates the standardized Henle unit values ​​of each voxel based on the patient's preoperative three-dimensional medical image data; the biomechanical property conversion module 120 converts the standardized Henle unit values ​​into bone biomechanical property values ​​based on a "Henle unit-biomechanical property" calibration model, generating a biomechanical property map; the map visualization module 160 displays the biomechanical property map using color mapping and overlays it with a three-dimensional bone model to obtain a fused visualization map; and the path planning and analysis module 130 automatically or interactively plans the bone entry path based on the biomechanical property map and surgical requirements, or based on the fused visualization map.

[0069] The system disclosed in this embodiment expands the navigation dimension of surgical robots from a single geometric space to a mechanical property space, enabling preoperative prediction and visualization of bone mechanical properties by constructing a mechanical property map. This improvement addresses the core deficiency of existing surgical robot navigation systems that rely solely on morphological information and cannot predict individual differences in bone mechanical properties. It intuitively presents bone mechanical performance information (which traditional imaging techniques like X-rays) by integrating a visualization map; it transforms complex and abstract mechanical concepts into intuitive and operable visual information, greatly reducing the cognitive burden on operators and improving the efficiency and accuracy of research and decision-making. Through the path planning and analysis module 130, it plans the bone entry path based on the mechanical property map, elevating bone entry path planning from "avoiding geometrically dangerous structures" to "anticipating mechanically difficult areas," realizing a shift from passive mechanical response to active planning. High-hardness areas on the planned path can be identified preoperatively, allowing for targeted strategies (such as pre-selecting a better path or adjusting operating parameters), improving the predictability and safety of the surgery, and reducing the risks of instrument deviation, tissue damage, and drill jamming due to unexpected encounters with hard bone.

[0070] In some embodiments, the image processing module 110 includes a data acquisition unit, a preprocessing unit, a bone tissue segmentation and extraction unit, and an image processing algorithm unit. The data acquisition unit acquires three-dimensional medical image data from a hospital PACS system or intraoperative equipment according to the DICOM standard protocol to form raw data. The preprocessing unit is connected to the data acquisition unit. The preprocessing unit performs anisotropic diffusion filtering on the raw data to reduce image noise while preserving bone edge information. The bone tissue segmentation and extraction unit is connected to the preprocessing unit. The bone tissue segmentation and extraction unit processes the preprocessed data using a region-growing-based bone tissue segmentation algorithm to extract the raw Henlein unit value for each voxel. The image processing algorithm unit is connected to the bone tissue segmentation and extraction unit. The image processing algorithm unit performs grayscale normalization on the raw Henlein unit values ​​of all extracted voxels to obtain standardized Henlein unit values ​​for all voxels, eliminating systematic errors caused by differences in different scanning equipment, thereby ensuring the accuracy and consistency of the extracted Henlein unit values.

[0071] Specifically, anisotropic diffusion filtering simply filters irregular pixels and identifies and preserves edges. The bone tissue segmentation and extraction unit uses a region-growing-based bone tissue segmentation algorithm to process the preprocessed 3D image data. Its core objective is to accurately identify and extract all regions belonging to the skeleton from a complete 3D image containing various tissues such as skin, muscle, fat, internal organs, and bone. It then measures the raw Henlein unit (HU) value of each voxel in the target bone region, providing the necessary data foundation for generating a mechanical property map. The algorithm determines whether each voxel in the 3D image should be included in the bone region based on specific rules and gradually aggregates them into the target region. A common method is to use the average HU value of the current region as a benchmark, including adjacent voxels within a certain threshold range into the bone region. For example, if the average HU value of the current bone region is 500, and the threshold is set to 100, then adjacent voxels with HU values ​​between 400 and 600 will be absorbed. However, this method has limitations when dealing with conditions such as osteolytic lesions of the spine, as the HU value of the cancellous bone in the lesion area may be significantly lower than normal levels, making it easy for the algorithm to exclude these lesions. In contrast, another method uses a higher HU threshold (e.g., greater than 700 HU) for initial segmentation, prioritizing the extraction of high-density cortical bone regions. Then, geometric morphological operations are used to fill in minute gaps caused by blood vessels and other structures, thus constructing a continuous and complete encapsulating shell. Finally, all voxels within this shell, regardless of their HU value, are identified as skeletal regions, thereby achieving effective extraction of all skeletal structures in the image and measurement of Henlein unit values. The image processing algorithm unit uses Python and calls the SimpleITK library to implement 3D image processing.

[0072] In some embodiments, the three-dimensional medical imaging data is CT or CBCT imaging data.

[0073] In some embodiments, the mechanical property value is bone stiffness or elastic modulus.

[0074] In one embodiment, the surgical requirements are input by the operator and are selected from bone tissue sampling, pedicle screw placement, or vertebral body distraction.

[0075] In some embodiments, the map visualization module 160 uses a color-region mapping scheme to convert mechanical property values ​​into intuitive color displays. The map visualization module 160 includes a color-region mapping unit, a WebGL-based real-time rendering engine, a user interaction unit, and an interactive query unit. The color-region mapping unit converts the mechanical property values ​​obtained by the mechanical property conversion module 120 into color displays using a continuous gradient color scale from blue to red, generating a colored mechanical property map; where blue represents low hardness values ​​and red represents high hardness values, and a legend is provided. The WebGL-based real-time rendering engine overlays the colored mechanical property map with the 3D skeletal model and sends the overlaid result to the terminal interactive interface for display. The user interaction unit is communicatively connected to the terminal interactive interface and responds to user operations on the transparency adjustment slider control on the terminal interactive interface, adjusting the transparency of the mechanical property map to observe the correspondence between the skeletal structure and mechanical properties. The interactive query unit is communicatively connected to the terminal interactive interface. The interactive query unit is used to respond to the user's click operation on the terminal interactive interface, obtain the spatial coordinates of the clicked position, query the corresponding standardized Heinz unit value and predicted mechanical property value based on the coordinates, and send the query results to the terminal interactive interface for display.

[0076] Visualized color-coded mechanical property maps can better assist in lesion diagnosis. Currently, clinicians utilize the principle that different tissue densities correspond to different gray levels in 3D images such as CT scans. By observing the gray levels in 3D images, they can initially identify and screen bone diseases such as osteolytic lesions (bone appears darker in the image), osteoblastic lesions (bone appears whiter in the image), and osteoporosis (bone appears grayish overall in the image). The mechanical property map of this invention is an upgrade based on this approach. It transforms bone density information into more direct and useful mechanical property data (such as hardness and strength), which can more clearly assist operators in assessing bone condition. In short, it upgrades from the traditional "looking at density" to "measuring strength."

[0077] In one example, the query results are displayed as a floating information box on the terminal interface.

[0078] In this embodiment, the mechanical property-color mapping table is shown in the following table:

[0079]

[0080] As mentioned above, the path planning and analysis module 130 has two path planning modes: interactive planning mode and automatic system planning mode. These two planning modes are described in detail below.

[0081] In interactive planning mode, after the operator manually plans the bone entry route on the integrated visualization map, the path planning and analysis module 130 automatically analyzes the mechanical property values ​​of the areas traversed by the bone entry path and generates a mechanical risk assessment report. The mechanical risk assessment report includes at least the total path length, the expected average cutting force, the location and value of the peak cutting force, and the proportion of high hardness areas.

[0082] In the automatic planning mode, the path planning and analysis module 130 generates multiple recommended bone entry paths with different focuses on safety, mechanical load and surgical efficiency by minimizing the comprehensive cost function based on the mechanical property map and surgical requirements, and outputs them for the operator to select.

[0083] In some embodiments, the comprehensive cost function takes into account the safety, operability and surgical efficiency of the path, and is a weighted sum of safety cost, mechanical load cost and efficiency cost, and its expression is: Total cost = W1 × f1 + W2 × f2 + W3 × f3;

[0084] Among them, W1, W2, and W3 are weighting coefficients, and W1 + W2 + W3 = 1, which are used to adjust the relative importance of each factor;

[0085] f1 is the safety cost term, and its value is negatively correlated with the minimum distance from the planned path to the critical anatomical structure. That is, the greater the distance, the better the safety.

[0086] f2 is the mechanical load cost term, and its value is positively correlated with the predicted cutting force of the area traversed by the planned path; the higher the mechanical load, the higher the surgical risk, and the higher the mechanical load cost term.

[0087] f3 is the efficiency cost term, and its value is negatively correlated with the theoretical average feed rate of the planned path under mechanical load; the slower the theoretical average feed rate, the lower the efficiency.

[0088] The path planning and analysis module 130 generates multiple recommended bone entry paths with different focuses, including the safest path, the most efficient path, and the balanced path, which the operator can choose according to the actual situation during the operation.

[0089] In some embodiments, the weighting coefficients W1, W2, and W3 can be adaptively configured according to the target surgical environment. The target surgical environment includes at least the patient condition, the type of surgery, and the surgical instruments.

[0090] When the surgical environment is a procedure with extremely high safety requirements, W1 should be configured as 0.4 ~ 0.6 and W2 as 0.2 ~ 0.4.

[0091] When the surgical environment is a routine surgical procedure, the configurations of W1, W2, and W3 are all within the range of 0.2 to 0.4.

[0092] When the surgical environment prioritizes efficiency, the W3 configuration should be 0.4 to 0.6.

[0093] An example of a procedure with extremely high safety requirements is complex spinal surgery adjacent to the spinal cord. A specific example of a routine procedure could be a standard approach with good bone quality. A specific example of a procedure prioritizing efficiency could be a bone biopsy.

[0094] In one embodiment, whether interactively planned or automatically planned by the system, for any planned bone entry path, the path planning and analysis module 130 is also configured to synchronously generate a predicted cutting force curve for that path based on the mechanical property values ​​of the area traversed by the bone entry path in the mechanical property map, and to visually provide the predicted cutting force curve to the operator. This curve will serve as a reference for real-time monitoring during the operation. The predicted cutting force curve represents the change in predicted cutting force with depth of feed as the surgical instrument 000 advances along the bone entry path.

[0095] This embodiment visualizes the predicted cutting force curve, making the invisible internal mechanical environment of the bone predictable, allowing the operator to have a preliminary understanding and psychological expectation of the mechanical situation to be faced before the operation begins.

[0096] In some embodiments, the process by which the path planning and analysis module 130 generates a predicted cutting force curve includes the following steps:

[0097] S110. Perform dense three-dimensional spatial sampling along the bone entry path; wherein, the sampling interval can be set according to the accuracy requirements, preferably 0.1mm to 0.5mm, to ensure the accuracy of the final predicted curve;

[0098] S120. For each path sampling point, query its skeletal mechanical property value in the mechanical property map;

[0099] S130. Using a pre-calibrated mechanical property conversion model, the mechanical property values ​​are converted into the predicted cutting force at the sampling point of the path;

[0100] S140. Based on the feed depth and predicted cutting force of all path sampling points, a smooth and continuous predicted cutting force curve is generated through data fitting.

[0101] The mechanical property conversion model establishes a mapping relationship between bone mechanical property values ​​and cutting forces. Considering the different specifications and models of surgical instruments 000 and the different operating parameters (e.g., rotational speed and feed rate) during use, standardized cutting experiments are conducted on bone samples of different hardness to generate a unique mechanical property conversion model for each specific specification and model of surgical instrument 000. When using a specific specification and model of surgical instrument 000 during surgery, selecting the specification and model of that surgical instrument 000 in the system will invoke the corresponding mechanical property conversion model. In one embodiment of the mechanical property conversion model, the mechanical property conversion model is expressed as:

[0102] Predicted cutting force = f(bone stiffness, instrument specifications, operating parameters)

[0103] Where f is a transformation function, the specific parameters of which are determined by standardized cutting experiments on bone samples of different hardness using surgical instruments of specific specifications and models 000, and by regression analysis based on the experimental results.

[0104] In some embodiments, data fitting is performed as spline interpolation to transform discrete prediction points into a continuous, smooth, and physically consistent prediction curve.

[0105] To ensure intraoperative safety, the system establishes a complete real-time monitoring and early warning mechanism. Accordingly, the system also includes a real-time monitoring and comparison module 140 and an early warning and / or control module 150. The real-time monitoring and comparison module 140 acquires real-time cutting force data of the surgical instrument 000 during its feed along the bone entry path and compares it in real-time with data from the predicted cutting force curve at the same location. The early warning and / or control module 150 executes at least one intraoperative guidance operation; the intraoperative guidance operation includes, but is not limited to: providing the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve formed from the real-time cutting force data and the predicted cutting force curve; outputting early warning information based on the comparison results; and adaptively adjusting the feed speed and / or rotation speed of the surgical instrument 000 through the surgical instrument power system 200, combining the current position mechanical property value with the comparison results. In this embodiment, the same feed depth is typically set at the same location; the current position mechanical property value refers to the mechanical parameter value corresponding to the current position of the surgical instrument 000 on the mechanical property map.

[0106] In some embodiments, the system integrates a force sensor 180 and a navigation and positioning device 190. The force sensor 180 is used to collect real-time cutting force data of the surgical instrument 000, and the navigation and positioning device 190 is used to simultaneously calculate the real-time spatial pose of the surgical instrument 000. The force sensor 180 and the navigation and positioning device 190 are respectively communicatively connected to the real-time monitoring and comparison module 140, so that the data from both are incorporated into the real-time monitoring and comparison module 140. The force sensor 180 is preferably a six-dimensional force and torque sensor, and the navigation and positioning device 190 is an optical navigation system.

[0107] The real-time monitoring and comparison module 140 performs fusion analysis on multi-source data and focuses on visualizing the cutting force and real-time pose deviation that are of most clinical concern. Other data (such as motor operating parameters in the power system) are recorded and monitored in the background. Once the early warning and / or control module 150 detects a real-time cutting force deviation, it will immediately trigger multi-level early warnings and execute preset safety intervention measures.

[0108] In some embodiments, the force sensor 180 uses an ATI Mini45, and the navigation and positioning device 190 uses an NDIPolaris. A multi-threaded data acquisition program was developed, with the data sampling frequency of the force sensor 180 set to 1000Hz and the data sampling frequency of the navigation and positioning device 190 set to 100Hz.

[0109] Multi-threaded data acquisition includes: the six-dimensional forces and dynamic torques acting on the surgical instrument 000; motor parameters providing rotation and displacement functions for the surgical instrument 000 (e.g., current, temperature, voltage of DC motors and stepper motors used to detect motor operating status); and the position angle deviation of the surgical instrument 000 monitored in real time by the NDI. This acquisition program needs to be set in the real-time monitoring and early warning module to collect and monitor motor and sensor data in real time. The position angle deviation of the surgical instrument 000 is synchronously interpreted by the navigation and positioning device 190.

[0110] The following is a detailed description of the intraoperative guidance operations performed by the early warning and / or control module 150.

[0111] In one embodiment, the early warning and / or control module 150 provides the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve and the predicted cutting force curve formed from real-time cutting force data.

[0112] Specifically, the early warning and / or control module 150 uses the predicted cutting force curve as a reference baseline and overlays it with the real-time cutting force data of the surgical instrument 000 on the terminal interface for the operator to compare intuitively. In some embodiments, the predicted cutting force curve and the measured cutting force curve are displayed in real time using a dual Y-axis chart.

[0113] This embodiment uses the predicted cutting force curve as a reference baseline on the terminal interactive interface to overlay and present the real-time cutting force data of the surgical instrument 000, upgrading the surgery from a "qualitative" operation that relies on the operator's personal experience and vague senses to a "quantitative and visual" precise operation that is guided by objective data and provides real-time feedback.

[0114] In one embodiment, the early warning and / or control module 150 outputs early warning information based on the comparison results, including: triggering an alarm of the corresponding level when the real-time cutting force data continues for a specified duration exceeding a multiple of the corresponding predicted value by different multiple thresholds; wherein, the specified duration is determined based on the data transmission frequency, and the specified duration is set to be greater than the transmission period of a single data point, so that the specified duration can cover multiple consecutive data points, thereby eliminating abnormal value interference with a duration equal to the transmission period of a single data point. For example, if the data transmission frequency F = 1000Hz, and the number of consecutive data points selected to be covered is 10, then the specified duration is 10 / 1000Hz, i.e., 0.01s; if the data transmission frequency F = 100Hz, and the number of consecutive data points selected to be covered is 10, then the specified duration is 10 / 100Hz, i.e., 0.1s.

[0115] This embodiment employs an early warning method based on "duration" and "tiered thresholds." Only when the deviation between the measured force and the predicted force is confirmed to be persistent and escalating will different levels of alerts be issued to the operator based on the degree of deviation. Compared to the simple threshold method of "warning at the first sign of deviation," this is more accurate and effective, improving the operator's experience.

[0116] In some embodiments, the alarm mode of the warning and / or control module 150 is selected from one or a combination of visual alarms and auditory alarms.

[0117] In some embodiments, different levels of alerts include:

[0118] Level 1 Alarm: When the real-time cutting force data continuously exceeds the first preset multiple of the predicted value but does not exceed the second preset multiple of the predicted value, the yellow visual warning of the surgical instrument 000 is triggered, and the warning icon is displayed on the terminal interface;

[0119] Level 2 Alarm: When the real-time cutting force data continuously exceeds a second preset multiple of the predicted value, a red visual warning is triggered, accompanied by an audible alarm, and a pause command is sent to the surgical instrument power system 200. Upon receiving the pause command, the surgical instrument power system 200, such as a DC motor or stepper motor, will stop working, thereby pausing the movement of the surgical instrument 000.

[0120] This embodiment avoids abrupt shutdowns through tiered processing: when real-time cutting force data consistently exceeds a first preset multiple of the predicted value but not a second preset multiple, the system considers it low-risk and only issues a warning. If the operator subjectively judges the risk to be high, they can manually pause the operation (by pressing the emergency stop button, releasing the foot pedal, etc.). When real-time cutting force data consistently exceeds a second preset multiple of the predicted value for a specified duration, it may be due to path deviation, instrument failure, or other reasons, making continued surgery risky. In this case, the surgery is immediately paused for examination. Low-risk warnings avoid unnecessary interruptions caused by brief, acceptable increases in force (e.g., penetrating a known cortical bone), improving operational continuity and reducing the overall surgical time extension caused by false alarms. High-risk forced pauses avoid risks from path deviation or instrument failure, balancing safety and efficiency.

[0121] This embodiment uses a graded alarm based on the degree of deviation between real-time cutting force data and chip force prediction to help operators distinguish between "expected high resistance" and "unexpected path deviation or machine failure," making it easier for operators to understand and make decisions.

[0122] In some embodiments, the first preset multiple is 130%, and the second preset multiple is 180%.

[0123] In some embodiments, in order to achieve visual and audible alarms, the surgical instrument 000 is equipped with a yellow indicator light, a red indicator light, and an alarm sound generating device. The yellow indicator light, the red indicator light, and the alarm sound generating device are respectively connected to the warning and / or control module 150, so that the warning and / or control module 150 illuminates the yellow indicator light during a first-level alarm and illuminates the red indicator light and the alarm sound generating device during a second-level alarm.

[0124] In some embodiments, the warning and / or control module 150 adaptively adjusts the feed speed and / or rotation speed of the surgical instrument 000 through the surgical instrument power system 200 according to preset graded control rules. These preset graded control rules include: when it is predicted that the surgical instrument 000 will enter a high-hardness region marked on the mechanical property map, the warning and / or control module 150 reduces the feed speed and increases the rotation speed of the surgical instrument 000 through the surgical instrument power system 200. Reducing the feed speed aims to decrease the amount of bone cut per unit time, thereby controlling the magnitude of the cutting force; when the real-time cutting force data continuously exceeds a preset multiple of the predicted value for a specified duration (this is usually caused by instrument failure, actual hardness exceeding the predicted value, etc.), the warning and / or control module 150 further reduces the feed speed of the surgical instrument 000 through the surgical instrument power system 200 to maintain stable cutting force and safety; when it is predicted that the surgical instrument 000 will enter a low-hardness region marked on the mechanical property map, the warning and / or control module 150 increases the feed speed of the surgical instrument 000 through the surgical instrument power system 200 to improve the overall efficiency of the surgery while ensuring safety.

[0125] This embodiment designs an adaptive control strategy based on a mechanical property map and real-time force feedback, enabling dynamic and intelligent adjustment of the speed and rotation speed of the surgical instrument 000. This improvement addresses the insufficient adaptability of conventional bone entry methods to complex and time-varying biological tissues, achieving a balanced optimization of safety and efficiency. The system no longer simply and passively responds to changes in force, but can adjust in advance when high resistance is predicted, making the operation smooth and controllable, protecting tissues and instruments; and appropriately increase speed when low resistance is predicted, thereby improving overall surgical efficiency while ensuring safety.

[0126] Preferably, the above control logic can be implemented through parameterization. Specifically, the system pre-stores standard feed rate and standard rotation speed as benchmarks. The definition of "high hardness region" and "low hardness region" can be achieved by setting thresholds based on the numerical distribution of the mechanical property map.

[0127] As an example of parameter configuration: when it is predicted that the surgical instrument 000 will enter a high hardness region, the feed rate of the surgical instrument 000 will be adjusted to a first specific proportion of the feed rate standard value, and the rotational speed of the surgical instrument 000 will be increased to a second specific proportion of the rotational speed standard value; when the real-time cutting force continues for a specified duration to exceed a third specific proportion of the predicted value, the feed rate of the surgical instrument 000 will be further adjusted to a fourth specific proportion of the feed rate standard value; when it is predicted that the surgical instrument 000 will enter a low hardness region, the feed rate can be increased to a fifth specific proportion of the feed rate standard value.

[0128] In some embodiments, the first specific ratio ranges from 50% to 70%, the second specific ratio ranges from 110% to 130%, the third specific ratio ranges from 140% to 160%, the fourth specific ratio ranges from 20% to 40%, and the fifth specific ratio ranges from 110% to 130%. Preferably, the first specific ratio ranges from 60%, the second specific ratio ranges from 120%, the third specific ratio ranges from 150%, the fourth specific ratio ranges from 30%, and the fifth specific ratio ranges from 120%.

[0129] Accordingly, in one specific embodiment, the region with Heinz unit values ​​between 500 HU and 700 HU is defined as the high hardness region, and the region with Heinz unit values ​​between 250 HU and 500 HU is defined as the low hardness region.

[0130] Those skilled in the art will understand that the specific values ​​of the aforementioned thresholds and adjustment ratios can be configured and optimized according to different surgical types (such as the different force sensitivities of vertebroplasty and pedicle screw placement), different instrument models, and the operator's individual preferences. The core of this invention lies in the control logic of "prediction-comparison-adaptive adjustment," rather than specific parameter values.

[0131] In some embodiments, after clinical deployment, to enable system self-evolution, the system further includes a data archiving and model iteration module 170, used to anonymize and store equipment information, dosage settings, predicted cutting force curves, and measured cutting force curves for each surgery, while complying with ethical and privacy regulations, and to periodically retrain and optimize the "Henness unit-mechanical property" calibration model using the accumulated data. Intraoperative mechanical feedback data includes, but is not limited to, real-time cutting force data of the surgical instrument 000; mechanical property prediction data includes, but is not limited to, the predicted cutting force curve of the bone entry path.

[0132] This embodiment constructs a closed-loop learning system of "data acquisition - model optimization - performance improvement" by setting up a data archiving and model iteration module 170. With the continuous accumulation of clinical applications, this system can continuously enhance its predictive accuracy, ensuring long-term leading performance and added value.

[0133] The following section uses bone stiffness as an example to illustrate the process of establishing the "Henness unit-mechanical property" calibration model before and after clinical application.

[0134] The process of establishing a preclinical calibration model includes the following steps:

[0135] S210. Acquire standard density phantoms, such as QRM's CT calibration phantom and multiple biological bone samples covering the target standardized Henle unit range.

[0136] Specifically, standard density phantoms, used to determine calibration values, are reference materials of known density, such as water, air, and equivalent materials of bone of different densities. Since patients of different ages, bone qualities, and surgical sites may be encountered during surgery, a large number of biological bone samples of varying densities are needed to simulate the diverse bone densities that may be encountered clinically. Furthermore, the Henle unit range for hardness conversion of these simulated biological bone samples should cover as many values ​​as possible within the target range, for example, 100, 120, 140, 160... within the 100-900 HU range. Pig bones raised under specific conditions can be used to meet the target standardized Henle unit range for hardness conversion.

[0137] S220. On at least three different mainstream CT devices, standard density phantoms and biological bone samples were scanned under different tube voltages and tube currents to establish device-specific Henle unit calibration curves.

[0138] Even with the same scanning parameters, CT scanners from different mainstream brands often produce different Heinz unit values. Therefore, we need to establish a device-specific Heinz unit calibration curve for each CT scanner to eliminate this inter-device bias and ultimately improve the accuracy of prediction results.

[0139] The process of obtaining the Heinz unit correction curve includes: using a standard phantom containing various materials with known densities (such as water, bone, and equivalent materials), measuring the measured Heinz unit values ​​(hereinafter referred to as measured HU values) of each material region in the phantom on the scanned image, establishing a correspondence between these values ​​and their known physical densities, generating a "measured HU-density" correction curve through linear regression fitting, and then converting the standard phantom density to the standard HU value (i.e., the standardized Heinz units mentioned above) to obtain the "measured HU-standard HU" correction curve, i.e., the "Heinz unit correction curve". The X-axis represents the measured HU value, and the Y-axis represents the corrected standard HU value. The measured HU value of the patient's image is substituted into the conversion. The curve needs to be verified by periodically scanning the standard phantom.

[0140] S230. Use a standard material testing machine to perform mechanical tests on the above samples to obtain their true mechanical property parameters for verification and optimization of the calibration model.

[0141] S240. Based on the standardized Heinz unit values ​​of biological bone samples and their corresponding actual mechanical property parameters, establish an initial mathematical relationship model between Heinz units and mechanical properties through multiple regression analysis, namely the initial "Henz unit-mechanical model" calibration model.

[0142] In some embodiments, the scikit-learn library of Python is used to train a random forest regression model. The input features include standardized Henness unit values ​​and their statistical features, and the output is the predicted mechanical property value.

[0143] A calibration model is constructed using a data-driven approach based on machine learning. In a preferred embodiment, a random forest regression model is trained using Python's scikit-learn library. Input features include not only the raw Henlein unit values ​​of the sampling points, but more importantly, various statistical features encompassing their neighborhoods to capture information about the spatial structure and stiffness of the bone. These features primarily include first-order statistical features, gradient features for identifying sharp transition zones between cortical and cancellous bone, and texture features that quantify the reticular structure of cancellous bone. The first-order statistical features can be the mean, standard deviation, skewness, or kurtosis of the Henlein unit values ​​within a 3×3×3 voxel block centered on the sampling point. The mean reflects local average density, the standard deviation reflects bone homogeneity, and skewness and kurtosis indicate the microstructural distribution of trabecular bone. Gradient features refer to the Henlein unit gradient values ​​of the sampling point in different directions. Texture features can be contrast, correlation, or entropy calculated based on the gray-level co-occurrence matrix. The model outputs predicted mechanical property values, such as the elastic modulus (GPa) or stiffness (MPa) of the bone. During the model training phase, the biological bone sample dataset is randomly divided into a training set and a test set (e.g., 70%–30%). The random forest model is trained using the training set, and the best performance is obtained through hyperparameter tuning.

[0144] S250. The initial calibration model is validated, and it can only be put into clinical use after it has met the predetermined comprehensive performance standards.

[0145] Specifically, this step involves validation including stability and generalization ability validation, statistical performance indicator threshold compliance validation, and clinical discriminant validity validation. Stability and generalization ability validation uses K-fold cross-validation to validate the initial calibrated model, assessing its stability and generalization ability across different data subsets to ensure it does not overfit the training data. Statistical performance indicator threshold compliance validation requires the calibrated model to simultaneously meet the following three predetermined thresholds on an independent test set: a coefficient of determination greater than or equal to 0.80, indicating the model can explain more than 80% of the variation in mechanical properties; a root mean square error less than 0.5 GPa, ensuring the average error between the predicted elastic modulus and the true value is within clinically acceptable limits; and a mean absolute percentage error less than 15%, further controlling prediction accuracy from a relative error perspective. Clinical discriminant validity validation requires the calibrated model to successfully distinguish between clinically significant different bone quality grades (e.g., normal bone, osteopenia, osteoporosis), with its classification accuracy showing significant statistical consistency with expert evaluation results. Only when the calibration model meets all three of the above requirements is it considered to have reliable predictive capabilities and be able to generate and provide accurate mechanical property maps to support clinical surgery.

[0146] The clinical data acquisition and archiving process includes the following steps: During clinical surgery, the system's data archiving and model iteration module 170 anonymizes and stores at least the following data: the patient's original CT parameters, predicted mechanical properties generated based on the calibration model, real-time cutting force data, motion parameters of the surgical instruments 000, and surgical outcome evaluation. This data is stored in a hierarchical structure; the original data is stored on the hospital's local server, while the anonymized feature data is uploaded to the cloud for calibration model iteration.

[0147] The iterative update process for the calibration model after clinical use includes the following steps: periodically retraining the "Henness unit-mechanical property" calibration model using accumulated data; updating the model parameters using an incremental learning algorithm; generating an updated calibration model; and maintaining model version management to continuously improve system performance. The updated calibration model is then deployed to the clinical system for subsequent surgeries. It should be understood that the update cycle can be flexibly set according to the rate of clinical data production (e.g., quarterly, semi-annually).

[0148] This embodiment also provides an intelligent guidance method for orthopedic surgical robots based on a mechanical property map, which includes preoperative steps, intraoperative steps, and postoperative steps.

[0149] The preoperative steps include:

[0150] The standardized Henlein unit values ​​for each voxel were calculated based on the patient's preoperative three-dimensional medical imaging data.

[0151] Based on the "Henle unit-mechanical property" calibration model, standardized Henle unit values ​​are converted into skeletal mechanical property values ​​to generate a mechanical property map;

[0152] The mechanical property map is displayed using color mapping and overlaid with the 3D skeletal model to obtain a fused visualization map.

[0153] Based on the mechanical property map and surgical requirements, the bone entry path is automatically or interactively planned based on the fused visualization map, and the predicted cutting force curve is generated by combining the mechanical property values ​​of the tissue areas traversed by the bone entry path in the mechanical property map.

[0154] Intraoperative steps include:

[0155] The real-time cutting force data of the surgical instrument 000 during its feeding along the bone entry path is obtained and compared in real time with the data at the same position of the predicted cutting force curve;

[0156] Perform at least one intraoperative guided operation; the intraoperative guided operation includes, but is not limited to: providing the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve and the predicted cutting force curve formed by the real-time cutting force data; outputting early warning information based on the comparison results; and adaptively adjusting the feed speed and / or rotation speed of the surgical instrument 000 through the surgical instrument power system 200 in combination with the current position mechanical property value and the comparison results, so as to realize the closed-loop linkage guidance of preoperative mechanical prediction and intraoperative mechanical feedback;

[0157] Postoperative steps include:

[0158] Under the premise of complying with ethical and privacy regulations, the data archiving and model iteration module 170 anonymizes and stores the equipment information, dosage settings, predicted cutting force curves and measured cutting force curves for each surgery, and uses the accumulated data to periodically retrain and optimize the "Henness unit-mechanical property" calibration model.

[0159] This invention provides an intelligent guidance system and method for orthopedic surgical robots based on mechanical property maps. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. An intelligent guidance system for orthopedic surgical robots based on a mechanical property map, characterized in that, include: The image processing module (110) is used to calculate the standardized Henle unit values ​​of each voxel based on the patient's preoperative three-dimensional medical image data. The mechanical property conversion module (120) is used to convert the standardized Heinz unit value into bone mechanical property value based on the "Henz unit-mechanical property" calibration model, and generate a mechanical property map; the mechanical property value is bone stiffness or elastic modulus. The map visualization module (160) is used to display the mechanical property map in a color mapping manner and overlay it with the skeletal 3D model to obtain a fused visualization map. And a path planning and analysis module (130) for automatically or interactively planning the bone entry path based on the mechanical property map and surgical requirements or based on the fused visualization map.

2. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to claim 1, characterized in that, The path planning and analysis module (130) is also configured to synchronously generate a predicted cutting force curve for the path based on the mechanical property values ​​of the area traversed by the bone entry path in the mechanical property map, and to provide the predicted cutting force curve to the operator in a visual manner.

3. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to claim 2, characterized in that, Also includes: The real-time monitoring and comparison module (140) is used to acquire real-time cutting force data of the surgical instrument (000) during the feeding process along the bone entry path and compare it with the data at the same position of the predicted cutting force curve in real time. And an early warning and / or control module (150) for performing at least one intraoperative guidance operation; the intraoperative guidance operation includes, but is not limited to: providing the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve formed by the real-time cutting force data and the predicted cutting force curve; outputting early warning information based on the comparison results; and adaptively adjusting the feed speed and / or rotation speed of the surgical instrument (000) through the surgical instrument power system (200) in combination with the current position mechanical property value and the comparison results.

4. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to claim 3, characterized in that, The output of early warning information based on the comparison results includes: When the real-time cutting force data continues for a specified duration for a number of times exceeding the corresponding predicted value and reaching different multiple thresholds, an alarm of the corresponding level is triggered; The duration of the specified duration is determined based on the data transmission frequency, and the duration of the specified duration is set to be greater than the transmission period of a single data point.

5. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to claim 4, characterized in that, The different levels of alerts include: Level 1 Alarm: When the real-time cutting force data continuously exceeds the first preset multiple of the predicted value but does not exceed the second preset multiple of the predicted value, a yellow visual warning is triggered, and a warning icon is displayed on the terminal interface; Level 2 alarm: When the real-time cutting force data continuously exceeds the second preset multiple of the predicted value, a red visual warning is triggered and accompanied by an audible alarm, and a pause command is sent to the surgical instrument power system (200).

6. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to claim 3, characterized in that, The warning and / or control module (150) adaptively adjusts the feed speed and / or rotation speed of the surgical instrument (000) through the surgical instrument power system (200) according to the preset hierarchical control rules; The preset hierarchical control rules include: When it is predicted that the device will enter a high-hardness region, the warning and / or control module (150) reduces the feed speed of the surgical instrument (000) and increases the rotation speed of the surgical instrument (000) through the surgical instrument power system (200). When the real-time cutting force data continues for a specified duration exceeding a preset multiple of the predicted value, the warning and / or control module (150) further reduces the feed speed of the surgical instrument (000) through the surgical instrument power system (200); When it is predicted that the instrument will enter a low hardness region, the warning and / or control module (150) increases the feed speed of the surgical instrument (000) through the surgical instrument power system (200).

7. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to any one of claims 1 to 3, characterized in that, The map visualization module (160) is also configured to provide a transparency adjustment function and a point-and-click query function for adjusting the overlay ratio of the mechanical property map and the skeletal 3D model display.

8. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to any one of claims 1 to 3, characterized in that, The interactive planning of the bone entry path based on the fused visualization map is specifically as follows: after the operator manually plans the bone entry route on the fused visualization map, the path planning and analysis module (130) automatically analyzes the mechanical property values ​​of the area traversed by the bone entry path and generates a mechanical risk assessment report; the mechanical risk assessment report includes at least the total path length, the expected average cutting force, the location and value of the peak cutting force, and the proportion of high hardness areas. And / or, the automatic bone entry path planning based on the mechanical property map and surgical requirements specifically means that: the path planning and analysis module (130) generates multiple recommended bone entry paths with different focuses on safety, mechanical load and surgical efficiency by minimizing the comprehensive cost function according to the mechanical property map and surgical requirements, and outputs them for the operator to select.

9. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to claim 3, characterized in that, This system integrates: A force sensor (180) is used to collect real-time cutting force data of the surgical instrument (000); And a navigation and positioning device (190) for synchronously calculating the real-time spatial pose of the surgical instrument (000); The force sensor (180) and the navigation and positioning device (190) are respectively connected to the real-time monitoring and comparison module (140) for communication, so that the real-time monitoring and comparison module (140) can obtain the real-time cutting force data and real-time spatial pose of the surgical instrument (000) from the force sensor (180) and the navigation and positioning device (190) respectively.

10. The intelligent guidance system for orthopedic surgical robots based on mechanical property maps according to any one of claims 1 to 3, characterized in that, It also includes a data archiving and model iteration module (170), which, under the premise of complying with ethical and privacy regulations, anonymizes and stores the equipment information, dosage settings, mechanical property prediction data and intraoperative mechanical feedback data for each surgery, and uses the accumulated data to periodically retrain and optimize the "Henness unit-mechanical property" calibration model.

11. A method for intelligent guidance of orthopedic surgical robots based on mechanical property maps, characterized in that, include: The standardized Henlein unit values ​​for each voxel were calculated based on the patient's preoperative three-dimensional medical imaging data. Based on the "Henle unit-mechanical property" calibration model, standardized Henle unit values ​​are converted into skeletal mechanical property values ​​to generate a mechanical property map; The mechanical property map is displayed using a color mapping method and overlaid with the 3D skeletal model to obtain a fused visual map. Based on the mechanical property map and surgical requirements, the bone entry path is automatically or interactively planned based on the fused visualization map, and a predicted cutting force curve is generated by combining the mechanical property values ​​of the tissue areas traversed by the bone entry path in the mechanical property map. Acquire real-time cutting force data of the surgical instrument (000) during the feeding process along the bone entry path and compare it in real time with the data at the same position of the predicted cutting force curve; Perform at least one of the intraoperative guidance operations; the intraoperative guidance operations include, but are not limited to: providing the operator with a real-time, dynamic, and visual comparison of the measured cutting force curve formed by the real-time cutting force data and the predicted cutting force curve; outputting warning information based on the comparison results; and adaptively adjusting the feed speed and / or rotation speed of the surgical instrument (000) through the surgical instrument power system (200) in combination with the current position mechanical attribute value and the comparison results, so as to realize the closed-loop linkage guidance of preoperative mechanical prediction and intraoperative mechanical feedback.

Images