Techniques capable of determining planned pose of medical implant

Abstract

The present invention relates to techniques capable of determining a planned pose of a medical implant. A method capable of determining a planned pose of a medical implant based on medical image data is disclosed. The server provides constraint information to the surgical planning station, the constraint information indicating at least one constraint to be used by the surgical planning station to determine a planned pose of the medical implant. The planning station obtains medical image data of at least a portion of the patient's body that is to be implanted with the medical implant and determines a planned pose of the medical implant based on the at least one constraint and the medical image data. An apparatus, computer program and system are also provided.

Description

Technical Field

[0001] The present disclosure provides a method capable of determining a planned posture of a medical implant based on medical image data. Also provided are an apparatus, a computer program, and a system. Background Art

[0002] In various surgical scenarios, the posture of a medical implant to be implanted in a patient's body is planned before the operation is performed. The surgeon can manually perform such planning based on preoperative medical image data of the patient's body. For example, a computed tomography (CT) scan of the patient's spine can be obtained, and a two-dimensional digitally rendered radiograph (DRR) generated based on the CT scan can be provided to the surgeon. The surgeon can then define in these DRRs the planned posture of the pedicle screws to be implanted in the patient's vertebrae later. Of course, such pre-planning of postures can also be used to plan the posture of non-spinal implants.

[0003] Some solutions support the surgeon in the pose planning stage by providing suggestions for a planned pose of a given medical implant. For example, a pose planning algorithm can be deployed locally on a planning station and determine the planned pose based on the patient's medical image data and predefined constraint information. Referring to the above-mentioned spine example, the pose planning algorithm deployed on the planning station can determine the planned pose of the pedicle screw based on the DRR by using the predefined geometric properties of the pedicle screw (e.g., a predefined diameter and a predefined length).

[0004] In the current solution, the accuracy of the planned poses will remain constant regardless of the number of planned poses determined by the pose planning algorithm over time. Summary of the invention

[0005] What is needed is a technique that solves one or more of the above or other problems.

[0006] According to a first aspect, a method is provided that is capable of determining a planned pose of a medical implant based on medical image data. The method is performed by a surgical planning station. The method includes receiving constraint information from a server, the constraint information indicating at least one constraint to be used by the surgical planning station to determine the planned pose of the medical implant. The method also includes obtaining medical image data of at least a portion of a patient's body in which the medical implant is to be implanted. The method also includes determining the planned pose of the medical implant based on the at least one constraint and the medical image data.

[0007] According to a second aspect, a method is provided for determining a planned pose of a medical implant based on medical image data. The method is performed by a server. The method includes sending constraint information to a surgery planning station, the constraint information indicating at least one constraint to be used by the surgery planning station to determine a planned pose of the medical implant based on medical image data of at least a portion of a patient's body in which the medical implant is to be implanted.

[0008] The following applies to the method according to the first aspect and the method according to the second aspect.

[0009] The constraint information may indicate only the at least one constraint. The constraint information may be provided from the server to the planning station as part of a data set consisting of non-executable data. The data set may not include executable computer program code. The method may not include providing a computer program from the server to the planning station.

[0010] The constraint information and / or the at least one constraint may be associated with a type of medical implant (eg, defined by user input obtained by the planning station and / or server).The at least one constraint may define the type of medical implant.

[0011] At least one constraint can limit one or more geometric properties of the medical implant. One or more geometric properties can include the size of the medical implant, the shape of the medical implant, the volume of the medical implant, and / or the profile of the medical implant. For example, the medical implant is a pedicle screw. In this case, the at least one constraint can limit the diameter and / or length of the pedicle screw.

[0012] Alternatively or additionally, the at least one constraint can define one or more spatial relationships between the medical implant and the anatomical unit of the patient's body. The one or more spatial relationships can include a predefined distance between at least a portion of the medical implant and the anatomical unit, a predefined position of at least a portion of the medical implant relative to the anatomical unit, and / or a predefined orientation of at least a portion of the medical implant relative to the anatomical unit. For example, the medical implant is a pedicle screw, and the anatomical unit is a vertebra. In this case, the at least one constraint can define the posture of the screw insertion trajectory relative to the vertebra.

[0013] The planning station may be equipped with a surgical planning program. Determining the planning pose may include processing the at least one constraint and the medical image data by the surgical planning program.

[0014] The medical image data may be hidden from the server, at least until the planned pose has been determined.Alternatively or additionally, (eg, any) information derived from the medical image data may be hidden from the server, at least until the planned pose has been determined.

[0015] The at least one constraint may be provided from the server to the surgical planning station preoperatively. The step of receiving constraint information according to the method of the first aspect may be performed preoperatively. The step of sending constraint information according to the method of the second aspect may be performed preoperatively. Optionally, one or more subsequent steps of the method may be performed intraoperatively. The method may not include a surgical step. Each of these methods may be referred to as a computer-implemented method.

[0016] Feedback information may be provided from the surgical planning station to the server. The method according to the first aspect may include sending feedback information to the server. The method according to the second aspect may include receiving feedback information from the planning station. The feedback information at least indicates: (i) one or more previously determined planning postures, (ii) at least one constraint for determining one or more previously determined planning postures, (iii) one or more previously determined and user-adjusted planning postures, and / or (iv) at least one constraint for determining one or more previously determined and user-adjusted planning postures, and the feedback information may be associated with one or more users (e.g., surgeons), one or more surgeries, one or more patients, one or more medical image types, and / or one or more hospitals.

[0017] At least one constraint to be used by the surgical planning station to determine the planned pose of the medical implant may be based on the feedback information. For example, the at least one constraint is determined by or was determined by the server. The method according to the second aspect may include, for example, determining the at least one constraint based on at least the feedback information.

[0018] For example, the at least one constraint is obtained or has been obtained as an output of a machine learning model, the output of which is trained using feedback information as training data. The machine learning model may be installed on a server or (for example, only) accessed by a server. The method according to the second aspect may include obtaining the at least one constraint from the machine learning model. The method according to the second aspect may include training the machine learning model based on feedback information. The method according to the second aspect may include obtaining feedback information from a plurality of planning stations and training the machine learning model based on the feedback information.

[0019] For example, federated learning is or was used to train a machine learning model. The method according to the first and / or second aspect may include using federated learning to train a machine learning model (eg, based on feedback information from multiple planning stations).

[0020] The method according to the first aspect may comprise triggering visualization of the determined planned posture, for example by superimposing an indication of the determined planned posture in a (e.g., two-dimensional or three-dimensional) rendering of the medical image data. Posture information indicating the determined planned posture may be sent to a server, for example as feedback information (e.g., part of the feedback information).

[0021] In a particular variation that is not currently reflected in the wording of the claims, the term "pose" as used herein would be replaced with "pose and / or size and / or shape." In this case, instead of or in addition to determining the planned pose of the medical implant, the planned size and / or planned shape of the medical implant could be determined.

[0022] According to a third aspect, a device is provided. The device is configured to perform the method according to the first aspect. In this case, the device can be configured as a planning station. Alternatively, the device is configured to perform the method according to the second aspect. In this case, the device can be configured as a server.

[0023] According to a fourth aspect, a computer program is provided. The computer program comprises instructions, which, when executed by a processor, cause the processor to perform a method according to the first aspect or according to the second aspect. The computer program may be carried by a carrier (e.g., a data stream, a signal wave, or a (e.g., non-transitory) computer-readable storage medium).

[0024] According to a fifth aspect, a system is provided. The system comprises a planning station configured to perform the method according to the first aspect. The system also comprises a server configured to perform the method according to the second aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Examples of the technology disclosed herein are described below with reference to the accompanying drawings, in which:

[0026] Figure 1 An exemplary system according to the present disclosure is shown;

[0027] Figure 2 A flowchart showing an exemplary method according to the present disclosure; and

[0028] Figure 3 An exemplary automated implant planning process according to the present disclosure is shown. DETAILED DESCRIPTION

[0029] Unless otherwise specified, reference numerals used hereinafter denote identical or similar structural or functional features. Where an example shows more than one instance of a given entity (which is denoted by reference numeral "X"), these instances may be referred to as "X" or "Xn", where n denotes a specific instance.

[0030] Figure 1An exemplary system 1000 according to the present disclosure is shown. System 1000 includes: a plurality of planning stations 100-1 to 100-3; a server 200, communicatively connected to each of the planning stations 100-1 to 100-3; and a database 300, communicatively connected to one or more of the planning stations 100-1 to 100-3. Each planning station 100 includes a processor 2 and a memory 4, which stores instructions that, when executed by the processor 2, configure the planning system 100 to operate in accordance with the manner disclosed herein. Server 200 also includes a processor 3 and a memory 5, which stores instructions that, when executed by the processor 3, configure the server 200 to operate in accordance with the manner disclosed herein. Each of the planning systems 100-1 to 100-3 also includes a display unit 6-1 to 6-3 configured to display visualization. Database 300 can store medical image data of a patient's body. The database 300 may be part of a picture archiving and communication system (PACS) (e.g., a PACS of a hospital) (e.g., in which the planning system 100 connected to the database 300 is also located). The database 300 may also be connected to or part of a medical image acquisition device (e.g., a magnetic resonance (MR) scanner or a computed tomography (CT) scanner), which may be configured to acquire medical image data.

[0031] The server 200 may be implemented as a single entity or a cloud computing platform service. On the other hand, the planning stations 100 are each configured as a local system, such as an independent system in one or more hospitals. The medical image data may be hidden from the server 200 to improve data protection. For example, the planning stations 100 may each be configured with a firewall that does not allow the server 200 to access the medical image data.

[0032] Figure 2 A flow chart of an exemplary method according to the present disclosure is shown. The method can be performed by the system 1000 or its components and can include optional steps indicated by dashed lines. One or more or all steps (e.g., at least steps 32 to 36) can be performed preoperatively.

[0033] In step 22, the planning station 100 obtains medical image data from the database 300. The medical image data includes one or more medical images of at least a portion of the patient's body where the medical implant will be implanted. The medical image data may include two-dimensional and / or three-dimensional image data, such as an MR scan, a CT scan, an X-ray image, or an ultrasound image of the portion of the patient's body. Each planning station 100-1 to 100-3 may obtain different medical image data. For example, the planning station 100-1 may obtain medical image data of a first patient, the planning station 100-2 may obtain medical image data of a second patient, and the planning station 100-3 may obtain medical image data of a third patient.

[0034] In optional step 24, the planning station 100 determines the planned pose of the implant to be implanted in the part of the patient's body based on the obtained medical image data. This step can be referred to as automatic implant planning. The planning station 100 can be equipped with a surgical planning program (e.g., stored on the memory 4) that processes at least the obtained medical image data to determine the planned pose.

[0035] Figure 3 An exemplary automatic implant planning process that can be performed as part of step 24 (eg, by a surgical planning program) is shown in FIG. 4 and includes sub-steps 45 , 46 , 48 , 50 .

[0036] The planning station 100 may obtain user input in substep 45 to define the type of medical implant to be implanted (e.g., bone screw, bone plate, spinal cage, etc.) and optionally the anatomical unit in which the medical implant is to be implanted (e.g., the name of a vertebra, such as "L4").

[0037] In sub-step 46, at least one anatomical unit in which the medical implant is to be implanted is segmented based on the medical image data. It will be clear to a person skilled in the art that various image segmentation techniques may be used in this context.

[0038] In sub-step 48, one or more landmarks that segment the anatomical unit are identified (eg, by comparing the segmented shape to a predefined template having pre-labeled landmarks).

[0039] In sub-step 50, a planned pose of the medical implant to be implanted is determined based on the segmented anatomical units and / or based on the identified landmarks. In this sub-step, the planned pose is determined using predefined constraint information. The predefined constraint information may be stored locally on the memory 4 (e.g., when a surgical planning program is installed on the planning station 100). Alternatively or additionally, the predefined constraint information may be defined (e.g., at least partially) by a user and / or obtained in sub-step 45. The predefined constraint information may indicate one or more predefined constraints to be used by the planning station 100 (e.g., the surgical planning program) to determine the planned pose, such as a predefined diameter and a predefined length of a pedicle screw and a predefined angle between a screw insertion trajectory of the pedicle screw and a vertebral end plate of a vertebra in which the pedicle screw is to be implanted.

[0040] In an optional step 26, display of a visualization on a display device 6 of the planning station 100 is triggered. The visualization at least indicates the planning pose determined in step 24. For example, a two-dimensional or three-dimensional rendering of one or more medical images included in the medical image data may be displayed, wherein a virtual medical implant conforming to the constraint information is superimposed on the rendering in the determined planning pose. Figure 1 As shown in FIG. Figure 1 , visualization 8 is displayed on display device 6-1 indicating the planned pose of pedicle screw 12 relative to vertebra 10 of a first patient, visualization 9 is displayed on display device 6-2 indicating the planned pose of pedicle screw 16 relative to vertebra 14 of a second patient, and visualization 11 is displayed on display device 6-3 indicating the planned pose of cage 20 relative to vertebra 18 of a third patient. Other variations of indicating the planned pose are also possible, for example by displaying the screw trajectory, the screw profile and / or the position of one or more predefined points (e.g., the screw tip) of the medical implant.

[0041] The planned posture determined in step 24 may not conform to the user's preferences. Therefore, the method may allow the user to change the determined posture and / or the medical implant whose posture is determined by providing appropriate user input. Therefore, in optional step 28, the user of the planning station 100 provides user input to the planning station 100, so that the planning station 100 (e.g., a surgical planning program) adjusts the planned posture determined in step 24 accordingly based on the user input. This in turn may result in adjusting at least one constraint used to determine the planned posture.

[0042] In optional step 30, the planning station 100 sends feedback information to the server 200. The feedback information may indicate one or more of the following: (i) a planned pose, (ii) predefined constraint information used to determine the planned pose in step 24, (iii) user input, and (iv) at least one adjusted constraint resulting from the adjusted planned pose. Alternatively or additionally, the feedback information sent in step 30 may indicate (i) one or more previously determined (e.g., as determined in step 24) planned poses, (ii) at least one (e.g., predefined) constraint used in determining the one or more previously determined planned poses, (iii) one or more previously determined and user adjusted (e.g., as adjusted in step 28) planned poses, and / or (iv) at least one (e.g., predefined) constraint used in determining the one or more previously determined and user adjusted planned poses. The feedback information may not include medical image data previously obtained by the planning station 100.

[0043] The server 200 may use feedback information obtained from one or more planning stations 100 for one or more surgeries and / or one or more patients and determine (e.g., updated and / or improved) constraint information based thereon. By way of example only, if the predefined constraint information indicates a preferred angle of 70° between the screw trajectory and the vertebral endplate, and the feedback information obtained from various planning stations 10 indicates an average preferred angle of 60°, it may be assumed that the predefined constraint information does not conform to the preferences of the users of the planning stations 10. Therefore, the constraint information may be updated by the server. To this end, the server 200 may train a machine learning model based on the feedback information in optional step 32. The machine learning model may be run on the server 200 or accessed by the server 200 (e.g., where it is provided by a cloud computing platform). The server 200 may at least input the feedback information as training data into the machine learning model to train it.

[0044] The server can also use federated learning to train the machine learning model. In this case, separate instances of the machine learning model can be trained separately using feedback information obtained from multiple planning stations 100, so that each instance is associated with a different planning station 100. Different instances can also be associated with different patients and / or different surgeries and / or different hospitals. The server 200 can then aggregate the trained instances to obtain a single trained machine learning model.

[0045] The trained machine learning model may be configured to provide (e.g., updated and / or improved) constraint information as output. Thus, in optional step 34, the server 200 obtains such constraint information from the trained machine learning model. The updated and / or improved constraint information may then be supplied to the planning station 100 to improve future planning results.

[0046] In step 36, the server 200 sends constraint information (e.g., the constraint information) to the planning station 100. The constraint information sent in step 36 may have been determined by the server 200, for example by using a trained machine learning model according to step 34. The constraint information sent in step 36 may be different from the predefined constraint information used in step 24. The constraint information sent in step 36 indicates at least one constraint to be used by the surgery planning station 100 (e.g., a surgery planning program) for determining a planned pose of the medical implant based on medical image data of at least a portion of a patient's body in which the medical implant is to be implanted. In the case where feedback information (e.g., the feedback information obtained in step 30) is used to train the machine learning model (e.g., in step 32), it can be said that the at least one constraint indicated by the constraint information is based on the feedback information.

[0047] The at least one constraint indicated by the constraint information sent in step 36 may define one or more geometric properties of the medical implant. For example, the one or more geometric properties may include the size of the medical implant, the shape of the medical implant, the volume of the medical implant, and / or the profile of the medical implant. If the medical implant is a pedicle screw, the at least one constraint defines the diameter and / or length of the pedicle screw.

[0048] Alternatively or additionally, at least one constraint indicated by the constraint information sent in step 36 may define one or more spatial relationships between the medical implant and an anatomical unit of the patient's body (e.g., in or on which the implant will be placed). For example, the one or more spatial relationships include a predefined distance between at least a portion of the medical implant and the anatomical unit, a predefined position of at least a portion of the medical implant relative to the anatomical unit, and / or a predefined orientation of at least a portion of the medical implant relative to the anatomical unit. If the medical implant is a pedicle screw and the anatomical unit is a vertebra, the at least one constraint may define a posture of the screw insertion trajectory relative to the vertebra.

[0049] In optional step 38, medical image data of at least a portion of the patient's body in which the medical implant is to be implanted is obtained. If desired, step 38 can also be avoided and the same medical image data already obtained in step 22 can be used. The medical image data that can be obtained in optional step 38 can be associated with a different patient, surgery, anatomical region and / or image acquisition time point than the medical image data obtained in step 22.

[0050] In step 40, a planned pose of the medical implant is determined based on the medical image data and further based on the (e.g., updated and / or improved) constraint information received in step 36. This differs from the determination of step 24 because, instead of using predefined constraint information, constraint information provided by the server 200 in step 36 is used. It can be said that the determination of the planned pose in step 40 is influenced by the server 200 because it relies on the (e.g., updated and / or improved) constraint information previously provided by the server 200. This provides the operator of the server (e.g., the manufacturer of the planning station 100) with the opportunity to improve the planning of the planning station 100 without having to roll out a new version of the surgical planning program and install it on each planning station 100. Compared to a complete surgical planning program including an executable computer program portion, this constraint information requires less storage space and can be sent using lower transmission resources.

[0051] In optional step 42, display of a visualization on the display unit 6 is triggered. The visualization may be configured similarly to the visualization described with reference to optional step 26, but based on the planned posture determined in step 40 instead of the planned posture determined in step 24 (and optionally, based on the medical image data obtained in step 38 instead of step 22). Step 40 may include one or more sub-steps 45 to 50, in which the constraint information obtained in step 36 is used, and optionally in sub-steps 46 and / or 48, the medical image data obtained in step 38 is used.

[0052] As described above with respect to optional step 28 , the user may disapprove of the planned gesture determined in step 40 and provide corresponding user input to initiate adjustment of the planned gesture in step 44 .

[0053] In optional step 46, the planning station 100 sends feedback information to the server 200. The feedback information may indicate one or more of the following: (i) the planned pose determined in step 40, (ii) (e.g., updated and / or improved) constraint information used to determine the planned pose in step 40, (iii) a user input initiating adjustment of the planned pose, and (iv) at least one adjusted constraint resulting from the adjusted planned pose. Alternatively or additionally, the feedback information sent in step 46 may indicate (i) one or more previously determined (e.g., as determined in steps 24 and / or 40), (ii) at least one (e.g., updated and / or improved) constraint used in determining the one or more previously determined planned poses, (iii) one or more previously determined and user-adjusted (e.g., as adjusted in step 44) planned poses, and / or (iv) at least one (e.g., updated and / or improved) constraint used in determining the one or more previously determined and user-adjusted planned poses. Again, here, the feedback information may not include medical image data previously obtained by the planning station 100.

[0054] The method may then continue with the server 200 determining or obtaining updated and / or improved constraint information based on the feedback obtained in step 46 and optionally further based on the feedback obtained in step 30. In this way, user adjustments to the planned posture determined based on the constraint information provided by the server may be fed back to the server to iteratively improve the constraint information, thereby producing a planned posture that better meets the user's requirements.

[0055] The technology disclosed herein will now be rephrased in other terms to explain the relevant teachings in more detail.

[0056] In general, planning the pose of a medical implant and / or its size is a time-consuming but important step in the navigation process of (eg, spinal) surgery. This functionality may be particularly advantageous in robot-assisted surgery, as the intended pose may need to be communicated to the surgical robot.

[0057] Automatic planning of implants can reduce the overall operation time and provide higher accuracy. This automatic planning process can be based on the detection of structures and anatomical landmarks in medical images (e.g., in steps 46 and 48). Machine learning can be used to segment anatomical units and / or identify prominent landmarks (e.g., in steps 46 and / or 48), such as pedicles, vertebral bodies, spinous processes, and transverse processes). Through the identified landmarks, the surgical planning program determines the planned posture of the desired implant.

[0058] In some solutions, fixed statistical models are used for automatic implantation planning, which relies on predefined constraint information. This solution is not perfect, and the surgeon may need to modify the planned posture to adapt to anatomical variations, previous operations or some potential vertebral anomalies. Therefore, also in order to provide better plans and improve overall planning accuracy, technology disclosed herein allows updating the statistical model of the automatic implantation plan, particularly predefined constraint information, based on feedback information (e.g., historical data of the adjustment made by the surgeon under determined implant posture).

[0059] It can be said that the present technology can improve the automatic implant plan based on feedback data recorded and saved on the server 200 (e.g., part of a cloud computing platform) without updating the surgical planning program installed on the planning station 1000. This can improve the accuracy of the determined posture without the need to redeploy the surgical planning program. The server 200 can, for example, use statistical and / or machine learning methods to calculate the improved weights of the constraints to be used to determine the planned posture of the implant. The improved weights can be fed as part of the updated constraint information in step 36. This in turn can improve the surgical planning program so that the subsequently planned posture provides higher clinical acceptance.

[0060] In other words, a conception of technology disclosed herein is to dynamically adapt the constraints for determining the planned posture of implant based on the feedback information collected for the adjustment of previously determined planned posture. This can be based on the changes made by the surgeon to the previously determined planned posture (for example, pedicle screw), for automatic implant posture plan to provide better constraints (for example, vertebra specific parameters, such as pedicle transverse angle, length and diameter). The surgeon can make changes to determined planned posture, to adapt anatomical deformation, previous surgery or some potential vertebrae abnormalities. Based on these changes, the server can use statistics and / or machine learning to calculate optimization weights as a part of updated constraint information, and these optimization algorithm weights can be fed to existing algorithms (for example, surgical planning program), to achieve higher accuracy, thereby speeding up the planning program.

[0061] The technology disclosed herein is not limited to screw plan optimization and spinal applications. It can be applied to other anatomical regions (e.g., shoulders, hips, or skulls), and can be used to plan the posture of implants. Similarly, in this case, improved implant plan postures and updated constraint information can be provided without changing the source code of the surgical planning program. That is, the surgical planning program can use (e.g., updated and / or improved) constraint information as parameter input for determining the planned posture in step 40. Further modifications and advantages of the technology disclosed herein will be apparent to those skilled in the art.

[0062] This disclosure also covers the following examples:

[0063] Example 1. A method capable of determining a planned pose of a medical implant based on medical image data, the method being performed by a surgical planning station and comprising:

[0064] receiving constraint information from a server, the constraint information indicating at least one constraint to be used by the surgical planning station to determine a planned pose of the medical implant;

[0065] obtaining medical image data of at least a portion of a patient's body in which the medical implant is to be implanted; and

[0066] A planned pose of the medical implant is determined based on the at least one constraint and the medical image data.

[0067] Example 2. A method capable of determining a planned pose of a medical implant based on medical image data, the method being performed by a server and comprising:

[0068] Constraint information is sent to the surgical planning station, the constraint information indicating at least one constraint to be used by the surgical planning station to determine a planned pose of the medical implant based on medical image data of at least a portion of the patient's body where the medical implant is to be implanted.

[0069] Example 3. The method according to Example 1 or 2, wherein:

[0070] The at least one constraint defines one or more geometric properties of the medical implant.

[0071] Example 4. The method according to Example 3, wherein:

[0072] The one or more geometric properties include a size of the medical implant, a shape of the medical implant, a volume of the medical implant, and / or a profile of the medical implant.

[0073] Example 5. The method according to Example 3 or 4, wherein:

[0074] The medical implant is a pedicle screw, and the at least one constraint defines a diameter and / or a length of the pedicle screw.

[0075] Example 6. The method according to any one of Examples 1 to 5, wherein:

[0076] The at least one constraint defines one or more spatial relationships between the medical implant and an anatomical element of the patient's body.

[0077] Example 7. The method according to Example 6, wherein

[0078] The one or more spatial relationships include a predefined distance between at least a portion of the medical implant and the anatomical element, a predefined position of at least a portion of the medical implant relative to the anatomical element, and / or a predefined orientation of at least a portion of the medical implant relative to the anatomical element.

[0079] Example 8. The method according to Example 6 or 7, wherein

[0080] The medical implant is a pedicle screw and the anatomical element is a vertebra, and wherein the at least one constraint defines a posture of the screw insertion trajectory relative to the vertebra.

[0081] Example 9. The method according to any one of Examples 1 to 8, wherein:

[0082] The planning station is equipped with a surgical planning program, wherein determining the planning pose comprises processing the at least one constraint and the medical image data by the surgical planning program.

[0083] Example 10. The method according to any one of Examples 1 to 9, wherein:

[0084] The medical image data, and optionally any information derived therefrom, are hidden from the server, at least until the planned pose has been determined.

[0085] Example 11. The method according to any one of Examples 1 to 10, wherein:

[0086] The at least one constraint is provided from the server to the surgical planning station preoperatively, wherein, optionally, one or more subsequent steps of the method are performed intraoperatively.

[0087] Example 12. The method according to any one of Examples 1 to 11, wherein:

[0088] providing feedback information from the surgical planning station to the server, the feedback information indicating at least: (i) one or more previously determined planned poses, (ii) at least one constraint used to determine the one or more previously determined planned poses, (iii) one or more previously determined and user-adjusted planned poses, and / or (iv) at least one constraint used to determine the one or more previously determined and user-adjusted planned poses,

[0089] Therein, at least one constraint to be used by the surgical planning station for determining a planned pose of the medical implant is based on the feedback information.

[0090] Example 13. The method according to any one of Examples 1 to 12, wherein:

[0091] The at least one constraint is or was determined by the server.

[0092] Example 14. The method according to Example 12 or 13, wherein

[0093] The at least one constraint is or was obtained as an output of a machine learning model, which is trained using the feedback information as training data.

[0094] Example 15. The method according to Example 14, wherein

[0095] Use or have used federated learning to train machine learning models.

[0096] Example 16. An apparatus configured to perform the method steps of the method described in any one of Examples 1 to 15.

[0097] Example 17. A computer program comprising instructions, which, when executed by a processor, causes the processor to perform the method according to any one of Examples 1 to 15, wherein the computer program is optionally carried by a carrier.

[0098] Example 18. A system comprising:

[0099] A planning station configured to perform at least the method described in Example 1; and

[0100] A server configured to perform at least the method described in Example 2.

Claims

1. A method capable of determining a planned pose of a medical implant based on medical image data, the method being performed by a surgical planning station and comprising: receiving constraint information from a server, the constraint information indicating at least one constraint to be used by the surgical planning station to determine a planned pose of a medical implant; obtaining medical image data of at least a portion of a patient's body in which the medical implant is to be implanted; as well as A planned pose of the medical implant is determined based on the at least one constraint and the medical image data.

2. The method according to claim 1, wherein: The at least one constraint defines one or more geometric properties of the medical implant.

3. The method according to claim 2, wherein: The one or more geometric properties include a size of the medical implant, a shape of the medical implant, a volume of the medical implant, and / or a profile of the medical implant.

4. The method according to claim 3, wherein: The medical implant is a pedicle screw, and the at least one constraint defines a diameter and / or a length of the pedicle screw.

5. The method according to claim 1, wherein: The at least one constraint defines one or more spatial relationships between the medical implant and an anatomical element of a patient's body.

6. The method according to claim 5, wherein: The one or more spatial relationships include a predefined distance between at least a portion of the medical implant and the anatomical unit, a predefined position of at least a portion of the medical implant relative to the anatomical unit, and / or a predefined orientation of at least a portion of the medical implant relative to the anatomical unit.

7. The method according to claim 5, wherein: The medical implant is a pedicle screw and the anatomical element is a vertebra, and wherein the at least one constraint defines a posture of a screw insertion trajectory relative to the vertebra.

8. The method according to claim 1, wherein: The planning station is equipped with a surgical planning program, wherein determining the planning pose comprises processing the at least one constraint and the medical image data by the surgical planning program.

9. The method according to claim 1, wherein: The medical image data, and optionally any information derived therefrom, are hidden from the server, at least until the planned pose has been determined.

10. The method according to claim 1, wherein: The at least one constraint is provided from the server to the surgical planning station preoperatively, wherein, optionally, one or more subsequent steps of the method are performed intraoperatively.

11. The method according to claim 1, wherein: providing feedback information from the surgical planning station to the server, the feedback information indicating at least: (i) one or more previously determined planned poses, (ii) at least one constraint used to determine the one or more previously determined planned poses, (iii) one or more previously determined and user-adjusted planned poses, and / or (iv) at least one constraint used to determine the one or more previously determined and user-adjusted planned poses, Therein, at least one constraint to be used by the surgical planning station to determine a planned pose of the medical implant is based on the feedback information.

12. The method according to claim 1, wherein: The at least one constraint is or was determined by the server.

13. The method according to claim 1, wherein: The feedback information is provided from the surgical planning station to the server, wherein the at least one constraint is or was obtained as an output of a machine learning model, the machine learning model being trained using the feedback information as training data.

14. The method according to claim 13, wherein: The machine learning model is or was trained using federated learning.

15. A surgical planning station, configured to: receiving constraint information from a server, the constraint information indicating at least one constraint to be used by the surgical planning station to determine a planned pose of a medical implant; obtaining medical image data of at least a portion of a patient's body in which the medical implant is to be implanted; as well as A planned pose of the medical implant is determined based on the at least one constraint and the medical image data.

16. A method capable of determining a planned pose of a medical implant based on medical image data, the method being performed by a server and comprising: Constraint information is sent to a surgical planning station, the constraint information indicating at least one constraint to be used by the surgical planning station to determine a planned pose of the medical implant based on medical image data of at least a portion of a patient's body where the medical implant is to be implanted.

17. The method according to claim 16, wherein: The medical image data, and optionally any information derived therefrom, are hidden from the server, at least until the planned pose has been determined.

18. The method according to claim 16, wherein: providing feedback information from the surgical planning station to the server, the feedback information indicating at least: (i) one or more previously determined planned poses, (ii) at least one constraint used to determine the one or more previously determined planned poses, (iii) one or more previously determined and user-adjusted planned poses, and / or (iv) at least one constraint used to determine the one or more previously determined and user-adjusted planned poses, Therein, at least one constraint to be used by the surgical planning station to determine a planned pose of the medical implant is based on the feedback information.

19. The method according to claim 16, wherein: The at least one constraint is or was determined by the server.

20. The method according to claim 16, wherein: The feedback information is provided from the surgical planning station to the server, wherein the at least one constraint is or was obtained as an output of a machine learning model, the machine learning model being trained using the feedback information as training data.

Images