Simulator for transforming supine three-dimensional geometry to standing position measurements

Abstract

Systems and methods for transforming supine three-dimensional geometry to standing position measurements. An example system may include an electronic processor. The electronic processor is configured to retrieve a scan of a spinal region in a supine position and process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans. The electronic processor is configured to receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region and generates, based on the plurality of weight¬ bearing measurements for the spinal region, a weight-bearing two-dimensional scan of the spinal region. The electronic processor is configured to provide as training data to a second machine learning model the scan of the spinal region in a supine position and the two-dimensional scan of the spinal region.

Description

SIMULATOR FOR TRANSFORMING SUPINE THREE-DIMENSIONAL GEOMETRY TO STANDINGPOSITION MEASUREMENTSFIELD

[0001] This application relates generally to the field of medical imaging systems.BACKGROUND

[0002] Predictive and recommendation models are used to plan and perform robotic-aided spinal surgeries. Such models are trained using patient electronic medical records, X-ray images, computed tomography scans, and other planning information related to the surgery to be performed.BRIEF DESCRIPTION OF THE DRAWINGS

[0003] The accompanying figures, where like reference numerals refer to identical or functionally similar elements throughout the separate views, together with the detailed description below, are incorporated in and form part of the specification, and serve to further illustrate embodiments, examples, aspects, and features of concepts that include the claimed subject matter and explain various principles and advantages of those embodiments, examples, aspects, and features.

[0004] FIG. 1 is a block diagram illustrating a medical computing system according to various examples.

[0005] FIG. 2 illustrates a flowchart of a method performed by the system of FIG. 1 according to various examples.

[0006] FIG. 3 illustrates a flowchart of a method for training a machine learning model to estimate spinal geometries according to various examples.

[0007] Skilled artisans will appreciate that elements in the figures are illustrated for simplicity and clarity and have not necessarily been drawn to scale. For example, the dimensions of some of the elements in the figures may be exaggerated relative to other elements to help to improve understanding of examples, aspects, and features illustrated.

[0008] In some instances, the apparatus and method components have been represented where appropriate by conventional symbols in the drawings, showing only those specific details that are pertinent to understanding the of various embodiments, examples, aspects, and features so as not to obscure the disclosure with details that will be readily apparent to those of ordinary skill in the art having the benefit of the description herein.DETAILED DESCRIPTION

[0009] Predictive and recommendation models are used by surgeons when planning and performing robotics assisted spinal surgery. Such models are trained using patient electronic medical records, X-ray images, computed tomography (CT) scans, and other planning information related to the surgery to be performed. These models are used to predict targets defined by geometric measures of the spine in the standing position. Targets include Adjacent Level Disease (ALD), proximal junctional kyphosis (PJK), and measures related to spinal geometry alignment (e.g., disc height, disc angle, spondylolisthesis, and vertebral end plate orientation). It is also desirable that the planning and execution phases of spinal surgeries, whether robotic assisted or freehand, should be performed in a proper standing CT.

[0010] Geometric measures of the spine in the standing position are typically taken using two- dimensional X-Ray (2D XR). An advantage of 2D XR is that it is typically done in the standing position, which represents well the global alignment of the patient in a weight bearing and balanced pose. However, 2D XR may be prone to unknown relative camera-patient poses, which can decrease the accuracy of the measurements. Another disadvantage of 2D XR is its limited accuracy when measuring exact spinal geometry parameters (e.g., disc height, spondylolisthesis grades, vertebral end plate orientation, etc.). Three-dimensional imaging (e.g., CT and magnetic resonance imaging (MRI)) scans improve accuracy, but these scans are taken in the supine position, which lacks relevant global alignment information (as opposed to the standing position).

[0011] To address these problems, embodiments and aspects presented herein provide machine learning models that transform supine three-dimensional spinal measurements to standing two- dimensional measurements. Such models, among other things, may be used to estimate accurate weight-bearing spinal geometry from supine CT scans. These models combine the advantages of two-dimensional and three-dimensional modalities and defines 3D-based criteria for geometric complication prediction.

[0012] Examples and aspects provided herein may be used to improve the performance of predictive spine analytics models. Predictive spine analytics models are trained on datasets of patients’ preoperative and postoperative data. Such models are used to predict postoperative outcomes using preoperative and planning data as inputs. Examples provided herein can provide enhanced predicted measurement quality be transforming the measurements from supine to weight-bearing position. Additionally, the predictive capabilities of such models are limited by, among other things, the quality of the input data. Examples provided herein can improve input data by filling in missing data (e.g., where a patient has a preoperative X-Ray and a postoperative CT scan) by automatically converting between CT and X-Ray measurements.

[0013] Using examples and aspects presented herein, spinal surgery planning may be improved. For example, surgical planning may be done based on a simulated weight bearing 3D scan produced from 3D scans in a supine position to achieve a desired alignment correction. Execution of spinal surgeries may also be improved. Using the examples presented herein, a surgeon can verify weightbearing alignment as the procedure is performed, allowing for corrections during the procedure, if necessary.

[0014] By estimating accurate weight-bearing spinal geometry from supine CT scan data, models provided herein may also be used my medical imaging systems to, among other things, product a simulated standing position (weight-bearing) CT scan from a supine CT scan. Such simulated CT scans may be used for surgical planning and improving spine analytics models, as noted above, for diagnostic purposes, etc.

[0015] In some aspects, the techniques described herein relate to a medical system. The system includes an electronic processor. The electronic processor is configured to retrieve a scan of a spinal region in a supine position. The electronic processor is configured to process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans. The electronic processor is configured to receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region. The electronic processor is configured to generate, based on the plurality of weight-bearing measurements for the spinal region, a weightbearing two-dimensional scan of the spinal region. The electronic processor is configured to generate a surgical plan for the spinal region based on the weight-bearing two-dimensional scan of the spinal region.

[0016] In some aspects, the techniques described herein relate to a method for training a machine learning model to estimate weight-bearing spinal geometry from supine scan data. The method includes generating, from a plurality of standing spinal X-Ray images, a plurality of three- dimensional standing images, wherein each of the plurality of three-dimensional standing images is derived from one of the plurality of spinal X-Ray images. The method includes associating each of a plurality of supine spinal scans with one of the plurality of three-dimensional standing images.The method includes providing as input data to the machine learning model the plurality of supine spinal scans. The method includes providing as ground truth data for the machine learning model the plurality of three-dimensional standing images.

[0017] In some aspects, the techniques described herein relate to a medical imaging system. The system includes a display and an electronic processor coupled to the display. The electronic processor is configured to retrieve a scan of a spinal region in a supine position. The electronic processor is configured to process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans. The electronic processor is configured to receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region. The electronic processor is configured to generate, based on the plurality of weightbearing measurements for the spinal region, a weight-bearing two-dimensional scan of the spinal region. The electronic processor is configured to control the display to present the weight-bearing two-dimensional scan of the spinal region.

[0018] In some aspects, the techniques described herein relate to a medical system. The system includes an electronic processor. The electronic processor is configured to retrieve a scan of a spinal region in a supine position. The electronic processor is configured to process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans. The electronic processor is configured to receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region. The electronic processor is configured to generate, based on the plurality of weight-bearing measurements for the spinal region, a weightbearing two-dimensional scan of the spinal region. The electronic processor is configured to provide as training data to a second machine learning model the scan of the spinal region in a supine position and the two-dimensional scan of the spinal region.

[0019] Specific embodiments of the present disclosure are now described with reference to the figures, wherein like reference numbers indicate identical or functionally similar elements.

[0020] Before any examples are explained in detail, it is to be understood that the examples presented herein are not limited in their application to the details of construction and the arrangement of components set forth in the following description or illustrated in the following drawings. The examples are capable of other embodiments and of being practiced or of being carried out in various ways. For ease of description, the example systems presented herein may be illustrated with a single exemplar of each of its component parts. Some examples may not describeor illustrate all components of the systems. Other example embodiments may include more or fewer of each of the illustrated components, may combine some components, or may include additional or alternative components.

[0021] FIG. 1 is a block diagram of a medical system 100. The medical system 100 may be used to carry out medical imaging, robotic assisted surgery, and other functions, including aspects of one or more of the methods disclosed herein. The medical system 100 includes a computing device 102, which is communicatively coupled (e.g., via one or more wired or wireless connections) to one or more imaging devices 112, a robot 114, a navigation system 118, one or more sensors 126, a database 130, and / or a cloud (or another network) 134. Systems according to other embodiments of the present disclosure may comprise more or fewer components than the medical system 100. The computing device 102 includes an electronic processor 104, a memory 106, a communication interface 108, and a user interface 110. In some aspects, the computing device 102 may include more or fewer components than illustrated in the example.

[0022] The computing device 102 includes an electronic processor 104 (for example, a microprocessor, application specific integrated circuit, etc.), a memory 106, a communication interface 108, and a user interface 110. The electronic processor 104, the memory 106, the communication interface 108, and the user interface 110, as well as the other various modules (not illustrated) are coupled directly, by one or more control or data buses (e.g., the bus 140), or a combination thereof.

[0023] The memory 106 may be made up of one or more non-transitory computer-readable media and includes at least a program storage area and a data storage area. The program storage area and the data storage area can include combinations of several types of memory, such as read-only memory (“ROM”), random access memory (“RAM”) (for example, dynamic RAM (“DRAM’), synchronous DRAM (“SDRAM”), etc.), electrically erasable programmable read-only memory (“EEPROM’), flash memory, or any other suitable tangible, non-transitory memory for storing computer- readable data and / or instructions. The memory 106 may store information or data useful for completing, for example, any aspect of the methods described herein, or of any other methods. The memory 106 may store, for example, instructions and / or machine learning models that support one or more functions of the system 100. For instance, the memory 106 may store content (e.g., instructions and / or machine learning models) that, when executed by the electronic processor 104, enable image processing 120, sensor processing 122, and / or the spinal geometry estimator 124.Such content may, in some embodiments, be organized into one or more applications, modules, packages, layers, or engines.

[0024] The image processing 120 enables the electronic processor 104 to process image data of an image (received from, for example, the imaging device 112, a camera of the navigation system 118, or any imaging device) for the purpose of, for example, identifying information about an anatomical element in an image.

[0025] In some aspects, the image processing 120 enables the electronic processor 104 to produce, analyze, manipulate, and otherwise process medical images produced, for example, by the imaging devices 112. For example, in some aspects the image processing 120 may enable the system 100 to produce and / or process two-dimensional X-Ray images, computed tomography (CT) scans, magnetic resonance imaging (MRI) scans, and the like.

[0026] The sensor processing 122 enables the processor 104 to process sensor output data (received from for example, the one or more sensors 126) for the purpose of, for example, determining the location of the robotic arm 116 and / or the surgical tool 128. The sensor output may be received as signal(s) and may be processed by the electronic processor 104 using the sensor processing 122 to output data such as, for example, force data, acceleration data, pose data, time data, location data, etc.

[0027] The spinal geometry estimator 124 enables the processor 104 to estimate accurate weightbearing spinal geometry from supine imaging scans, as described herein. As described more particularly herein, the spinal geometry estimator 124 implements a machine learning model. Machine learning generally refers to the ability of a computer program to leam without being explicitly programmed. In some embodiments, a computer program (sometimes referred to as a learning engine) is configured to construct a model (for example, one or more algorithms) based on example inputs. Supervised learning involves presenting a computer program with example inputs and their desired (actual) outputs. The computer program is configured to learn a general rule (a model) that maps the inputs to the outputs in the training data. Machine learning may be performed using various types of methods and mechanisms. Example methods and mechanisms include decision tree learning (e.g., random forests), association rule learning, artificial neural networks, deep neural networks, inductive logic programming, support vector machines, clustering, Bayesian networks, reinforcement learning, representation learning, similarity and metric learning, sparse dictionary learning, and genetic algorithms. Using some or all of theseapproaches, a computer program may ingest, parse, and understand data and progressively refine models for data analytics, including image analytics. Once trained, the computer system may be referred to as an intelligent system, an artificial intelligence (Al) system, a cognitive system, or the like. Alternatively, or additionally, the memory 106 may store other types of content or data (e.g., other types of machine learning models) that can be processed by the electronic processor 104 to carry out the various method and features described herein.

[0028] Although various contents of memory 106 may be described as instructions, it should be appreciated that functionality described herein can be achieved through use of instructions, algorithms, and / or machine learning models. The data, algorithms, and / or instructions may cause the electronic processor 104 to manipulate data stored in the memory 106 and / or received from or via the imaging device 112, the robot 114, the database 130, the sensors 126, and / or the cloud 134.

[0029] The electronic processor 104 sends and receives information (for example, from the memory 106, the communication interface 108, and / or the user interface 110) and processes the information by executing one or more software instructions or modules, capable of being stored in the memory 106, or another non-transitory computer readable medium. The software can include firmware, one or more applications, program data, filters, rules, one or more program modules, and other executable instructions. The electronic processor 104 is configured to retrieve from the memory 106 and execute, among other things, software for performing techniques and methods as described herein, including the image processing 120, sensor processing 122, and spinal geometry estimator 124.

[0030] The communication interface 108 transmits and receives information from devices external to the computing device 102, for example, components of the medical system 100. The communication interface 108 receives input (for example, from the user interface 110), provides system output or a combination of both. The communication interface 108 may be used for receiving image data or other information from an external source (such as the imaging device 112, the robot 114, the navigation system 118, the sensors 126, the database 130, the cloud 134, and / or any other system or component not part of the medical system 100), and / or for transmitting instructions, images, or other information to an external system or device (e.g., another computing device, imaging device, robot, navigation system, sensors, database, cloud, and / or any other system or component not part of the medical system 100).

[0031] The communication interface 108 may include one or more wired interfaces (e.g., a USB port, an Ethernet port, etc.) and / or one or more wireless transceivers or interfaces (configured, for example, to transmit and / or receive information via one or more wireless communication protocols such as 802.11a / b / g / n, Bluetooth, NFC, ZigBee, and so forth). In some embodiments, the communication interface 108 may be useful for enabling the computing device 102 to communicate with one or more other electronic processors or computing devices, whether to reduce the time needed to accomplish a computing-intensive task or for any other reason.

[0032] The computing device 102 may also include one or more user interfaces 110. The user interface 110 may be or include a keyboard, mouse, trackball, monitor, television, screen, touchscreen, and / or any other device for receiving information from a user and / or for providing information to a user. The user interface 110 may be used, for example, to receive a user selection or other user input regarding any step of any method described herein. Notwithstanding the foregoing, any required input for any step of any method described herein may be generated automatically by the medical system 100 (e.g., by the electronic processor 104 or another component of the medical system 100) or received by the medical system 100 from a source external to the medical system 100. In some embodiments, the user interface 110 may be useful to allow a surgeon or other user to modify instructions to be executed by the electronic processor 104 according to one or more embodiments of the present disclosure, and / or to modify or adjust a setting of other information displayed on the user interface 110 or corresponding thereto.

[0033] Although the user interface 110 is shown as part of the computing device 102, in some embodiments, the computing device 102 may utilize a user interface 110 that is housed separately from one or more remaining components of the computing device 102. In some embodiments, the user interface 110 may be located proximate one or more other components of the computing device 102, while in other embodiments, the user interface 110 may be located remotely from one or more other components of the computing device 102.

[0034] It should be understood that although FIG. 1 illustrates only a single electronic processor 104, memory 106, communication interface 108, and user interface 110, alternative embodiments of the computing device 102 may include multiple electronic processors, memory modules, communication interfaces, and / or user interfaces. It should also be noted that the medical system 100 may include other computing devices, each including similar components as, and configured similarly to, the computing device 102. In some embodiments, portions of the computing device102 are implemented partially or entirely on a semiconductor chip (e.g., an application specific integrated circuit (ASIC), a field-programmable gate array (“FPGA”), and the like). Similarly, the various modules and controllers described herein may be implemented as individual controllers, as illustrated, or as components of a single controller. In some aspects, a combination of approaches may be used. In addition, any one of the image processing 120, sensor processing 122, and / or the spinal geometry estimator 124 may be implemented on a separate computing platform but be accessible and / or controllable by the computing device 102.

[0035] Continuing with other aspects of the medical system 100, the imaging device 112 may be operable to image anatomical feature(s) (e.g., a bone, veins, tissue, etc.) and / or other aspects of patient anatomy to yield image data (e.g., image data depicting or corresponding to a bone, veins, tissue, etc.). “Image data” as used herein refers to the data generated or captured by an imaging device 112, including in a machine-readable form, a graphical / visual form, and in any other form. In various examples, the image data may comprise data corresponding to an anatomical feature of a patient, or to a portion thereof. In some aspects, the image data is stored (e.g., in the database 130, the cloud 134, or both) in the Digital Imaging and Communications in Medicine (DICOM) format. The image data may be or comprise a preoperative image, an intraoperative image, a postoperative image, or an image taken independently of any surgical procedure. The imaging device 112 may be capable of taking a 2D image (e.g., an X-Ray) or a 3D image (e.g., a CT scan or an MRI scan) to yield the image data. The imaging device 112 may be or comprise, for example, an ultrasound scanner (which may comprise, for example, a physically separate transducer and receiver, or a single ultrasound transceiver), an 0-arm, a C-arm, a G-arm, or any other device utilizing X-ray-based imaging (e.g., a fluoroscope, a CT scanner, or other X-ray machine), a magnetic resonance imaging (MRI) scanner, an optical coherence tomography (OCT) scanner, an endoscope, a microscope, an optical camera, a thermographic camera (e.g., an infrared camera), a radar system (which may comprise, for example, a transmitter, a receiver, a processor, and one or more antennae), or any other imaging device 112 suitable for obtaining images of an anatomical feature of a patient.

[0036] The sensors 126 are configured to provide sensor output. The sensors 126 may include a position sensor, a proximity sensor, a magnetometer, or an accelerometer. In some embodiments, the sensors 126 may include a linear encoder, a rotary encoder, or an incremental encoder (e.g., positioned to sense movement or position of the robotic arm 116). Sensor output or output datafrom the sensors 126 may be provided to an electronic processor of the robot 114, to the electronic processor 104 of the computing device 102, and / or to the navigation system 118. Output data from the sensor(s) 126 may also be used to determine position information for the robot 114. It will be appreciated that in some embodiments, the sensors 126 may be a component separate from the robotic arm 116. In other embodiments, sensors 136 — which may be the same as or similar to the sensors 126 — may be integrated with the robot 114. The sensors 126 may enable the electronic processor 104 (or an electronic processor of the robot 114) to determine a precise pose in space of a robotic arm 116 (as well as any object or element held by or secured to the robotic arm). In other words, sensor output or output data from the sensors 126 may be used to calculate a position in space of the robotic arm 116 (and thus, the surgical tool 128) relative to one or more coordinate systems.

[0037] The robot 114 may be any surgical robot or part of any robotic assisted surgery system, either of which is capable of operating as described herein. The robot 114 may be or comprise, for example, the Mazor X™ Stealth Edition robotic guidance system, or any derivative thereof. The robot 114 may be configured to position the surgical tool 128 at one or more precise poses (e.g., position(s) and orientation(s)). The surgical tool 128 may be any tool capable of cutting, drilling, milling, and / or parting an anatomical element. The surgical tool 128 may be, in one example, a drill bit. In some embodiments, the robot 114 may be configured to rotate and / or advance the cutting tool 128 using, for example, one or more motors to rotate the surgical tool 128.

[0038] The robot 114 may additionally or alternatively be configured to manipulate any component (whether based on guidance from the navigation system 118 or not) to accomplish or to assist with a surgical task. In some embodiments, the robot 114 may be configured to hold and / or manipulate an anatomical element during or in connection with a surgical procedure. The robot 114 may comprise one or more robotic arms 116. In some embodiments, the robotic arm 116 may comprise a first robotic arm and a second robotic arm, though the robot 114 may comprise more than two robotic arms. In some embodiments, one or more of the robotic arms 116 may be used to hold and / or maneuver the surgical tool 128. Each robotic arm 116 may be positionable independently of the other robotic arm. The robotic arms 116 may be controlled in a single, shared coordinate space, or in separate coordinate spaces.

[0039] The robot 114, together with the robotic arm(s) 116, may have, for example, one, two, three, four, five, six, seven, or more degrees of freedom. Further, the robotic arm 116 may be positionedor positionable in any pose, plane, and / or focal point. The pose includes a position and an orientation. As a result, a surgical tool 128 or another object held by the robot 114 (or, more specifically, by the robotic arm 116) may be precisely positionable in one or more needed and specific positions and orientations.

[0040] In some embodiments, reference markers (e.g., navigation markers) may be placed on the robot 114 (including, e.g., on the robotic arm 116), the surgical tool 128, or any other object in the surgical space. The reference markers may be tracked by the navigation system 118, and the results of the tracking may be used by the robot 114 and / or by an operator of the medical system 100 or any component thereof.

[0041] The navigation system 118 provides navigation for a surgeon and / or a surgical robot during an operation. The navigation system 118 may be any now-known or future-developed navigation system, including, for example, the Medtronic StealthStation™ S8 surgical navigation system or any successor thereof. The navigation system 118 may include one or more cameras or other sensor(s) for tracking one or more reference markers, navigated trackers, or other objects within the operating room or other room in which some or all of the medical system 100 is located. The one or more cameras may be optical cameras, infrared cameras, or other cameras. In some embodiments, the navigation system 118 may comprise one or more electromagnetic sensors. In various embodiments, the navigation system 118 may be used to track a position and orientation (e.g., a pose) of the imaging device 112, the robot 114 and / or robotic arm 116, the surgical tool 128), and / or one or more other tools (or, more particularly, to track a pose of a navigated tracker attached, directly or indirectly, in fixed relation to the one or more of the foregoing). The navigation system 118 may include a display for displaying one or more images from an external source (e.g., the computing device 102, imaging device 112, or other source) or for displaying an image and / or video stream from the one or more cameras or other sensors of the navigation system 118. In some embodiments, the medical system 100 can operate without the use of the navigation system 118. The navigation system 118 may be configured to provide guidance to a surgeon or other user of the medical system 100 or a component thereof, to the robot 114, or to any other element of the medical system 100 regarding, for example, a pose of one or more anatomical elements, whether or not a tool is in the proper trajectory, and / or how to move a tool into the proper trajectory to carry out a surgical task according to a preoperative or other surgical plan.

[0042] In some implementations, the database 130 stores patient data associated with one or more patients. Patient data may include, for example, medical images for one or more patients created, for example, by the imaging devices 112 or loaded into the database 130 from another source or system. Patient data may also include an age associated with a patient, a body mass index (“BMI”) associated with a patient, a bone density associated with a patient, a sex associated with a patient, an implant history associated with a patient (for example, a date and a placement location associated with an implant such as a screw, a rod, a cage, a prosthesis, or the like), other aspects of a patient’s medical history (e.g., smoker / non-smoker, chronic and acute illnesses, and the like), or combinations of the foregoing. In some implementations, patient data is associated with a patient undergoing a surgical procedure wherein the systems and methods described herein are utilized. In some implementations, the patient data may be stored in the database 130 prior to a surgical procedure. In some implementations, the electronic processor 104 is configured to send to the database 130, a query requesting patient data associated with a unique patient identifier and receive, from the database 130, patient data associated with the unique patient identifier.

[0043] The database 130 may store information that correlates one coordinate system to another (e.g., one or more robotic coordinate systems to a patient coordinate system and / or to a navigation coordinate system). The database 130 may additionally or alternatively store, for example, one or more surgical plans (including, for example, pose information about a target and / or image information about a patient’s anatomy at and / or proximate the surgical site, for use by the robot 114, the navigation system 118, and / or a user of the computing device 102 or of the medical system 100); one or more images useful in connection with a surgery to be completed by or with the assistance of one or more other components of the medical system 100; and / or any other useful information.

[0044] The database 130 may be configured to provide any such information to the computing device 102 or to any other device of the medical system 100 or external to the medical system 100, whether directly or via the cloud 134. In some embodiments, the database 130 may be or comprise part of a hospital image storage system, such as a picture archiving and communication system (PACS), a health information system (HIS), and / or another system for collecting, storing, managing, and / or transmitting electronic medical records including image data.

[0045] The cloud 134 may be or represent the Internet or any other wide area network. The computing device 102 may be connected to the cloud 134 via the communication interface 108,using a wired connection, a wireless connection, or both. In some embodiments, the computing device 102 may communicate with the database 130 and / or an external device (e.g., a computing device) via the cloud 134.

[0046] The medical system 100 or similar systems may be used, for example, to carry out one or more aspects of the methods described herein. The medical system 100 or similar systems may also be used for other purposes.

[0047] FIG. 2 illustrates an example method 200 for operating the system of FIG. 1 to perform robotic surgery using estimated weight-bearing spinal geometries. Although the method 200 is described in conjunction with the medical system 100 as described herein, the method 200 could be used with other systems and devices. In addition, the method 200 may be modified or performed differently than the example provided. In particular, the method 200 is also applicable to surgeries using free hand surgical tools.

[0048] As an example, the method 200 is described as being performed by the computing device 102, and, in particular, the electronic processor 104. However, it should be understood that, in some examples, portions of the method 200 may be performed by other components of the medical system 100 or components outside of the system 100.

[0049] At block 202, the electronic processor 104 retrieves a scan of a spinal region (e.g., the lumbar portion) in a supine position. For example, the electronic processor 104 may retrieve a computed tomography scan from the database 130 (e.g., taken during preoperative evaluation). In some aspects, the electronic processor 104 may control imaging devices 112, sensors 126, or another medical scanning device coupled to the system 100 to perform a scan of the spinal region within the operating theater as part of a surgical operation. In another example, the electronic processor 104 may retrieve or produce a magnetic resonance imaging scan.

[0050] At block 204, the electronic processor 104 processes the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans. For example, the electronic processor 104 may execute the spinal geometry estimator 124, which implements a machine learning model trained using the method 300, described below. In some examples, the scan data is provided to the machine learning model. In some examples, the electronic processor 104 first segments the images of the spinal region into a plurality of spinal segments (e.g., identifying and segmenting each vertebra making up the spinal region in the scan). The electronic processor 104 generates, for each of the segmented spinal segments, a three-dimensional surfacemodel (e.g., depicting the vertebra). Using these surface models, the electronic processor 104 determines, for each three-dimensional surface model, at least one three-dimensional geometric feature (e.g., a disc volume, a spondylolisthesis vector / surface, a three-dimensional disc orientation, and the like). In some aspects, the electronic processor provides as input to the machine learning model the three-dimensional geometric features for each of the plurality of spinal segments.

[0051] At block 206, the electronic processor 104 receives, from the machine learning model, a plurality of weight-bearing measurements for the spinal region. Weight bearing measurements are measurements describing the geometry of the spinal region while in a standing position, with the effects of gravity and the patient’s body compressing the spine. Weight-bearing measurements include an anterior disc height, a posterior disc height, a disc angle, a spondylolisthesis grade, and a spinal end plate orientation. In some aspects, these measurements are two-dimensional measurements, some of which are derived from corresponding three-dimensional geometric features (determined at block 204). For example, the machine learning model may predict the anterior and posterior disc heights from the disc volume, the spondylolisthesis grade from the spondylolisthesis vector / surface, and the spinal end plate orientation from the three-dimensional disc orientation. In some aspects, a two-dimensional measurement may be derived from combinations of its corresponding three-dimensional geometric features combined with other three-dimensional geometric features for the spinal region, from combinations of two or more three-dimensional geometric features not directly corresponding to the two-dimensional measurement, or from analysis of all of the three-dimensional geometric features for the spinal region.

[0052] At block 208, the electronic processor 104 generates, based on the plurality of weightbearing measurements for the spinal segment, a weight-bearing two-dimensional scan of the spinal region. For example, the scan retrieved at block 202 may be modified using image processing techniques by applying the weight bearing measurements to the scan to produce a simulated two- dimensional X-Ray scan having dimensions correlating to the weight bearing measurements. In some examples, the electronic processor 104 controls a display (e.g., part of the user interface 110) to present the weight-bearing two-dimensional scan (e.g., for analysis by a radiologist or surgeon. Optionally, the electronic processor 104 may generate a three-dimensional simulated scan, whichis based on lateral and anterior-posterior measurement predictions. In some aspects, the axial rotation is missing, resulting in a simulated three-dimensional scan up to axial rotation.

[0053] At block 208, the electronic processor 104 generates a surgical plan for the spinal region based on the weight-bearing two-dimensional scan of the spinal region. For example, the electronic processor 104 may execute an existing planning routing using the weight-bearing two-dimensional scan as input (for example, with input into the planning by a surgeon).

[0054] In some aspects, the electronic processor 104 controls the robot 114 to execute an operation on the spinal region based on the surgical plan. For example, the electronic processor 104 may control the robotic arm 116 according to the surgical plan and / or a user input to remove a volume of bone from the spinal region with the surgical tool 128. A surgeon may also manipulate the surgical tool 128 directly to perform bone removal assisted by the robotic arm, according to the surgical plan.

[0055] As part of the surgical procedure, the electronic processor 104 may determine a result of the operation. For example, while the surgical tool is being operated, the electronic processor 104 tracks the movement of the surgical tool 128. By tracking the movements of the surgical tool 128, the electronic processor 104 can use knowledge of the movements and the state of the surgical tool 128 (e.g., the type of tool head and how it is operating) to generate data representing the volume of bone being removed. Based on this result, the electronic processor 104 may generate, with the machine learning model, at least one updated weight-bearing measurement based on the result. For example, where a volume a bone has been removed, the machine learning model may generate an updated disc height value. In some examples, the electronic processor 104 may, as the surgical procedure progressed, generate an updated surgical plan based on the updated weight-bearing measurements. In some examples, the electronic processor 104 may also take subsequent supine images of the spinal segment during the procedure and use the images to produce updated weightbearing two-dimensional scans, which may be presented for the surgeon, used to generate updated surgical plans, or both.

[0056] In some aspects, the electronic processor 104 may alternatively or additionally use the weight-bearing two-dimensional scans as input for a second machine learning model. For example, supine spinal scans and their associated two-dimensional weight bearing scans may be used to supplement date provided to train other machine learning models, as described below.

[0057] FIG. 3 illustrates an example method 300 for training a machine learning model to estimate weight-bearing spinal geometries from supine image data. Although the method 300 is described in conjunction with the medical system 100 as described herein, the method 300 could be used with other systems and devices. In addition, the method 300 may be modified or performed differently than the example provided.

[0058] As an example, the method 300 is described as being performed by the computing device 102, and, in particular, the electronic processor 104. However, it should be understood that, in some examples, portions of the method 300 may be performed by other components of the medical system 100 or components outside of the system 100.

[0059] In the example below, the machine learning model is a random forest regressor. In other examples, the machine learning model may be a convolutional neural network, a transformer network, or a support vector regression network. Other variations of the machine learning model are also possible without deviating from the disclosure presented herein.

[0060] At block 302, the method 300 includes generating, from a plurality of standing spinal X- Ray images, a plurality of three-dimensional standing images. The spinal X-Ray images are posterior-anterior and lateral views taken from a sample of patients, where both standing X-Ray images and supine CT images (or other three-dimensional supine images) are available. However, because the target is new (a three-dimensional weight bearing image), there is no ground truth to use in training the machine learning model. Accordingly, the method 300 includes generating the ground truth for training. Each of the plurality of three-dimensional standing images is derived from one of the plurality of spinal X-Ray images.

[0061] In some aspects, for each of the spinal X-Ray images, the electronic processor 104 segments a spine depicted in the X-Ray image into a plurality of vertebra. The electronic processor 104 performs a three-dimensional pose reconstruction to generate vertebra poses for the vertebra (e.g., using six degrees of freedom). In some examples, the electronic processor 104 uses Computer Vision (CV) algorithms for stereoscopic vision to generate the vertebra poses, using the lower most vertebra as the reference. To reconstruct the pose of each vertebra, at least three corresponding points are needed, preferably four or more. To accommodate this, one of the X-Ray images should be in oblique position to allow for all four corners from one side to be visible in the oblique image. Alternatively, additional corresponding features’ points could be detected on the images using computer vision techniques.

[0062] Alternatively, or in addition, poses may be estimated using the Mazor registration algorithm, which optimizes both lateral and anterior / posterior registration, to provide per vertebra 6 DOFs pose estimation.

[0063] To build the three-dimensional standing image, the vertebrae are stitched together, each according to its predicted pose, using a lowest vertebra of the plurality of vertebra as an anchor.

[0064] In some examples, the resulting three-dimensional standing images are simulated three- dimensional computed tomography scans and / or simulated magnetic resonance imaging scans.

[0065] At block 304, the method 300 associates each of a plurality of supine spinal scans with one of the plurality of three-dimensional standing images. For example, the database 130 may include patient records including both the X-Ray images and the supine spinal scans. Using this information, the electronic processor 104 is able to assign each of the generated three-dimensional standing images with its corresponding supine spinal scan. The supine spinal scans may be three- dimensional computed tomography scans, magnetic resonance imaging scans, or both.

[0066] At block 306, the method 300 provides as input data to the machine learning model the plurality of supine spinal scans. The supine spinal scans are the type of inputs, for which the trained machine learning model will generate weight-bearing geometric measurements.

[0067] At block 308, the method 300 provides as ground truth data for the machine learning model the plurality of three-dimensional standing images. In some aspects, a more granular ground truth is provided. For example, the electronic processor 104 may determine, for each of the plurality of three-dimensional standing images, a plurality of weight-bearing measurements (as described above) and include in the ground truth data the pluralities of weight-bearing measurements.

[0068] In some examples, the machine learning model is trained using cross validation. For example, the input data and the ground truth data may be spilt into a number of training data subsets and used to perform a k-fold cross validation.

[0069] In some examples, the machine learning model, in producing its output, normalizes disc and vertebral sizes (width, depth and height, divided by spinal length), and vertebral and disc slenderness ratios.

[0070] In some aspects, the machine learning models described herein may be used to improve other analytic machine learning models. For example, for CT based criteria, applying the transformation on the training data set [D, t] in three dimensions for complication or planning prediction (D =data features input, t= target value, D' and t' are the same after the transformation)results in [D1, t']. In other aspects, the machine learning models described herein may be used for data enrichment. For example, existing two-dimensional and three-dimensional features can be enriched with transformed features (D U D',t) to train with the original target t.

[0071] Similarly, the machine learning models described herein may be used for modality consolidation for database extension. For example, to define a dynamic target, images in two different points in time are compared. Those images must be of the same modality to be comparable. However, in practice available scans are often mixed (e.g., having an X-Ray at time tl but a CT at time t2). In addition, many medical databases have missing data which causes loss of important “journey” records for model training. Using the transformation model presented herein allows for completion of databases. In the example described, the CT instance is transformed to equivalent representation of an X-Ray instance, increasing training data significantly.

[0072] With regard to the processes, systems, methods, heuristics, etc. described herein, it should be understood that, although the steps of such processes, etc. have been described as occurring according to a certain ordered sequence, such processes could be practiced with the described steps performed in an order other than the order described herein. It further should be understood that certain steps could be performed simultaneously, that other steps could be added, or that certain steps described herein could be omitted. In other words, the descriptions of processes herein are provided for the purpose of illustrating certain implementations and should in no way be construed to limit the claims.

[0073] Accordingly, it is to be understood that the above description is intended to be illustrative and not restrictive. Many embodiments and applications other than the examples provided would be apparent upon reading the above description. The scope should be determined, not with reference to the above description, but should instead be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled. It is anticipated and intended that future developments will occur in the technologies discussed herein, and that the disclosed systems and methods will be incorporated into such future embodiments. In sum, it should be understood that the application is capable of modification and variation.

[0074] All terms used in the claims are intended to be given their broadest reasonable constructions and their ordinary meanings as understood by those knowledgeable in the technologies described herein unless an explicit indication to the contrary in made herein. In particular, use of the singulararticles such as “a,” “the,” “said,” et cetera, should be read to recite one or more of the indicated elements unless a claim recites an explicit limitation to the contrary.

[0075] Unless explicitly stated otherwise, each numerical value and range should be interpreted as being approximate as if the word “about” or “approximately” preceded the value or range.

[0076] Reference herein to “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the disclosure. The appearances of the phrase “in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments necessarily mutually exclusive of other embodiments. The same applies to the term “implementation.”

[0077] Unless otherwise specified herein, the use of the ordinal adjectives “first,” “second,” “third,” etc., to refer to an object of a plurality of like objects merely indicates that different instances of such like objects are being referred to, and is not intended to imply that the like objects so referred-to have to be in a corresponding order or sequence, either temporally, spatially, in ranking, or in any other manner.

[0078] Unless otherwise specified herein, in addition to its plain meaning, the conjunction “if’ may also or alternatively be construed to mean “when” or “upon” or “in response to determining” or “in response to detecting,” which construal may depend on the corresponding specific context. For example, the phrase “if it is determined” or “if [a stated condition] is detected” may be construed to mean “upon determining” or “in response to determining” or “upon detecting [the stated condition or event]” or “in response to detecting [the stated condition or event].”

[0079] Also, for purposes of this description, the terms “couple,” “coupling,” “coupled,” “connect,” “connecting,” or “connected” refer to any manner known in the art or later developed in which energy is allowed to be transferred between two or more elements, and the interposition of one or more additional elements is contemplated, although not required. Conversely, the terms “directly coupled,” “directly connected,” et cetera, imply the absence of such additional elements. The same type of distinction applies to the use of terms “attached” and “directly attached,” as applied to a description of a physical structure. For example, a relatively thin layer of adhesive or other suitable binder can be used to implement such “direct attachment” of the two corresponding components in such physical structure.

[0080] The described embodiments are to be considered in all respects as only illustrative and not restrictive. In particular, the scope of the disclosure is indicated by the appended claims rather than by the description and figures herein. All changes that come within the meaning and range of equivalency of the claims are to be embraced within their scope.

[0081] The functions of the various elements shown in the figures, including any functional blocks labeled as “processors” and / or “controllers,” may be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functions may be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be shared. Moreover, explicit use of the term “processor” or “controller” should not be construed to refer exclusively to hardware capable of executing software, and may implicitly include, without limitation, digital signal processor (DSP) hardware, network processor, application specific integrated circuit (ASIC), field programmable gate array (FPGA), read only memory (ROM) for storing software, random access memory (RAM), and nonvolatile storage. Other hardware, conventional and / or custom, may also be included. Similarly, any switches shown in the figures are conceptual only. Their function may be carried out through the operation of program logic, through dedicated logic, through the interaction of program control and dedicated logic, or even manually, the particular technique being selectable by the implementer as more specifically understood from the context.

[0082] As used in this application, the term “circuitry” may refer to one or more or all of the following: (a) hardware-only circuit implementations (such as implementations in only analog and / or digital circuitry); (b) combinations of hardware circuits and software, such as (as applicable): (i) a combination of analog and / or digital hardware circuit(s) with software / firmware and (ii) any portions of hardware processor(s) with software (including digital signal processor(s)), software, and memory(ies) that work together to cause an apparatus, such as a mobile phone or server, to perform various functions); and (c) hardware circuit(s) and or processor(s), such as a microprocessor(s) or a portion of a microprocessor(s), that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation.” This definition of circuitry applies to all uses of this term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers an implementation of merely a hardware circuit or processor (or multiple processors) or portion of a hardware circuit or processorand its (or their) accompanying software and / or firmware. The term circuitry also covers, for example and if applicable to the particular claim element, a baseband integrated circuit or processor integrated circuit for a mobile device or a similar integrated circuit in server, a cellular network device, or other computing or network device.

[0083] It should be appreciated by those of ordinary skill in the art that any block diagrams herein represent conceptual views of illustrative circuitry embodying the principles of the disclosure. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and so executed by a computer or processor, whether or not such computer or processor is explicitly shown.

[0084] It should be understood that although certain figures presented herein illustrate hardware and software located within particular devices, these depictions are for illustrative purposes only. In some embodiments, the illustrated components may be combined or divided into separate software, firmware, and / or hardware. For example, instead of being located within and performed by a single electronic processor, logic and processing may be distributed among multiple electronic processors. Regardless of how they are combined or divided, hardware and software components may be located on the same computing device or may be distributed among different computing devices connected by one or more networks or other suitable communication links.

[0085] The following paragraphs provide various Examples reciting examples and alternatives disclosed herein.

[0086] Example 1. A medical system, the system comprising: an electronic processor configured to: retrieve a scan of a spinal region in a supine position; process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans; receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region; generate, based on the plurality of weight-bearing measurements for the spinal region, a weightbearing two-dimensional scan of the spinal region; and generate a surgical plan for the spinal region based on the weight-bearing two-dimensional scan of the spinal region.

[0087] Example 2. The medical system of Example 1, further comprising: a surgical robot communicatively coupled to the electronic processor; wherein the electronic processor is further configured to: execute an operation on the spinal region based on the surgical plan; determine a result of the operation; generate, with the machine learning model, at least one updated weight-bearing measurement based on the result; and generate an updated surgical plan based on the at least one updated weight-bearing measurement.

[0088] Example 3. The medical system of any of Examples 1 and 2, further comprising: a display communicatively coupled to the electronic processor; wherein the electronic processor is further configured to: control the display to present the weight-bearing two-dimensional scan.

[0089] Example 4. The medical system of any of Examples 1-3, wherein the electronic processor is further configured to process the scan using a machine learning model by: segment the spinal region into a plurality of spinal segments; generate, for each of the plurality of spinal segments, a three-dimensional surface model; determine, for each three-dimensional surface model, at least one three-dimensional geometric feature; and provide as input to the machine learning model the at least one three-dimensional geometric feature for each of the plurality of spinal segments.

[0090] Example 5. The medical system of any of Examples 1-4, wherein the electronic processor is configured to train the machine learning model by: generating, from a plurality of standing spinal X-Ray images, a plurality of three-dimensional standing images, wherein each of the plurality of three-dimensional standing images is derived from one of the plurality of spinal X-Ray images; associating each of a plurality of supine spinal CT scans with one of the plurality of three- dimensional standing images; providing as input data to the machine learning model the plurality of supine spinal CT scans; and providing as ground truth data for the machine learning model the plurality of three-dimensional standing images.

[0091] Example 6. The medical system of Example 5, wherein the electronic processor is further configured to: generate the plurality of three-dimensional standing images by: for each of a plurality of spinal X-Ray images, segmenting a spine in the image into a plurality of vertebra; for each of the plurality of vertebra, performing a three-dimensional pose reconstruction to generate a plurality of vertebra poses; and stitching each vertebra of the plurality of vertebra according to its predicted pose, using a lowest vertebra of the plurality of vertebra as an anchor.

[0092] Example 7. The medical system of any of Examples 5 and 6, wherein the electronic processor is configured to train the machine learning model by: splitting the input data and the ground truth data into a plurality of training data subsets; and performing a k-fold cross validation based on the plurality of training data subsets.

[0093] Example 8. The medical system of any of Examples 5-7, wherein the plurality of spinal X- Ray images includes a plurality of posterior-anterior and lateral views. 1

[0094] Example 9. The medical system of any of Examples 1-8, wherein the plurality of weightbearing measurements includes an anterior disc height, a posterior disc height, a disc angle, a spondylolisthesis grade, and a spinal end plate orientation.

[0095] Example 10. The medical system of any of Examples 1-9, wherein the scan of the spinal region is one of a three-dimensional computed tomography scan and a magnetic resonance imaging scan.

[0096] Example 11. The medical system of any of Examples 1-10, wherein weight-bearing two- dimensional scan of the spinal region is a simulated two-dimensional computed tomography scan.

[0097] Example 12. A method for training a machine learning model to estimate weight-bearing spinal geometry from supine scan data by: generating, from a plurality of standing spinal X-Ray images, a plurality of three-dimensional standing images, wherein each of the plurality of three- dimensional standing images is derived from one of the plurality of spinal X-Ray images; associating each of a plurality of supine spinal scans with one of the plurality of three-dimensional standing images; providing as input data to the machine learning model the plurality of supine spinal scans; and providing as ground truth data for the machine learning model the plurality of three-dimensional standing images.

[0098] Example 13. The method of Example 12, further comprising: determining, for each of the plurality of three-dimensional standing images, a plurality of weight-bearing measurements; and including in the ground truth data the pluralities of weight-bearing measurements; wherein the plurality of weight-bearing measurements for each of the plurality of three-dimensional standing images includes at least one selected from the group consisting of an anterior disc height, a posterior disc height, a disc angle, a spondylolisthesis grade, and a spinal end plate orientation.

[0099] Example 14. The method of any of Examples 12 and 13, wherein generating the plurality of three-dimensional standing images includes: for each of a plurality of spinal X-Ray images, segmenting a spine in the image into a plurality of vertebra; for each of the plurality of vertebra, performing a three-dimensional pose reconstruction to generate a plurality of vertebra poses; and stitching each vertebra of the plurality of vertebra according to its predicted pose, using a lowest vertebra of the plurality of vertebra as an anchor.

[0100] Example 15. The method of any of Examples 12-14, further comprising: splitting the input data and the ground truth data into a plurality of training data subsets; and performing a k-fold cross validation based on the plurality of training data subsets.

[0101] Example 16. The method of any of Examples 12-15, wherein the plurality of spinal X-Ray images includes a plurality of posterior-anterior and lateral views.

[0102] Example 17. The method of any of Examples 12-16, wherein the plurality of supine spinal scans includes at least one selected from the group consisting of a three-dimensional computed tomography scan and a magnetic resonance imaging scan.

[0103] Example 18. The method of any of Examples 12-17, wherein the plurality of three- dimensional standing images includes at least one selected from the group consisting of a simulated three-dimensional computed tomography scan and a simulated magnetic resonance imaging scan.

[0104] Example 19. The method of any of Examples 12-18, wherein the machine learning model is one selected from the group consisting of a random forest regressor, a convolutional neural network, a transformer network, and a support vector regression network.

[0105] Example 20. A medical imaging system, the system comprising: a display; an electronic processor coupled to the display and configured to: retrieve a scan of a spinal region in a supine position; process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans; receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region; generate, based on the plurality of weightbearing measurements for the spinal region, a weight-bearing two-dimensional scan of the spinal region; and control the display to present the weight-bearing two-dimensional scan of the spinal region.

[0106] Example 21. A medical system, the system comprising: an electronic processor configured to: retrieve a scan of a spinal region in a supine position; process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans; receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region; generate, based on the plurality of weight-bearing measurements for the spinal region, a weightbearing two-dimensional scan of the spinal region; and provide as training data to a second machine learning model the scan of the spinal region in a supine position and the two-dimensional scan of the spinal region.

[0107] Various features and advantages of the examples and embodiments presented herein are set forth in the following claims.

Claims

CLAIMSWhat is claimed is:

1. A medical system (100), the system comprising: an electronic processor (104) configured to: retrieve a scan of a spinal region in a supine position; process the scan using a machine learning model trained to produce standing spinal measurements from supine spinal scans; receive, from the machine learning model, a plurality of weight-bearing measurements for the spinal region; generate, based on the plurality of weight-bearing measurements for the spinal region, a weight-bearing two-dimensional scan of the spinal region; and generate a surgical plan for the spinal region based on the weight-bearing two-dimensional scan of the spinal region.

2. The medical system (100) of claim 1, further comprising: a surgical robot (114) communicatively coupled to the electronic processor (104); wherein the electronic processor (104) is further configured to: execute an operation on the spinal region based on the surgical plan; determine a result of the operation; generate, with the machine learning model, at least one updated weight-bearing measurement based on the result; and generate an updated surgical plan based on the at least one updated weight-bearing measurement.

3. The medical system (100) of any of claims 1 and 2, further comprising: a display communicatively coupled to the electronic processor (104); wherein the electronic processor (104) is further configured to: control the display to present the weight-bearing two-dimensional scan.

4. The medical system (100) of any of claims 1-3, wherein the electronic processor (104) is further configured to process the scan using a machine learning model by: segment the spinal region into a plurality of spinal segments; generate, for each of the plurality of spinal segments, a three-dimensional surface model;determine, for each three-dimensional surface model, at least one three-dimensional geometric feature; and provide as input to the machine learning model the at least one three-dimensional geometric feature for each of the plurality of spinal segments.

5. The medical system (100) of any of claims 1-4, wherein the electronic processor (104) is configured to train the machine learning model by: generating, from a plurality of standing spinal X-Ray images, a plurality of three- dimensional standing images, wherein each of the plurality of three-dimensional standing images is derived from one of the plurality of spinal X-Ray images; associating each of a plurality of supine spinal CT scans with one of the plurality of three- dimensional standing images; providing as input data to the machine learning model the plurality of supine spinal CT scans; and providing as ground truth data for the machine learning model the plurality of three- dimensional standing images.

6. The medical system (100) of claim 5, wherein the electronic processor (104) is further configured to: generate the plurality of three-dimensional standing images by: for each of a plurality of spinal X-Ray images, segmenting a spine in the image into a plurality of vertebra; for each of the plurality of vertebra, performing a three-dimensional pose reconstruction to generate a plurality of vertebra poses; and stitching each vertebra of the plurality of vertebra according to its predicted pose, using a lowest vertebra of the plurality of vertebra as an anchor.

7. The medical system (100) of any of claims 5 and 6, wherein the electronic processor (104) is configured to train the machine learning model by: splitting the input data and the ground truth data into a plurality of training data subsets; and performing a k-fold cross validation based on the plurality of training data subsets.

8. The medical system (100) of any of claims 5-7, wherein the plurality of spinal X-Ray images includes a plurality of posterior-anterior and lateral views.

9. The medical system (100) of any of claims 1-8, wherein the plurality of weight-bearing measurements includes an anterior disc height, a posterior disc height, a disc angle, a spondylolisthesis grade, and a spinal end plate orientation.

10. The medical system (100) of any of claims 1-9, wherein the scan of the spinal region is one of a three-dimensional computed tomography scan and a magnetic resonance imaging scan.

11. The medical system (100) of any of claims 1-10, wherein weight-bearing two-dimensional scan of the spinal region is a simulated two-dimensional computed tomography scan.

12. A method (300) (200) for training a machine learning model to estimate weight-bearing spinal geometry from supine scan data by: generating, from a plurality of standing spinal X-Ray images, a plurality of three- dimensional standing images, wherein each of the plurality of three-dimensional standing images is derived from one of the plurality of spinal X-Ray images; associating each of a plurality of supine spinal scans with one of the plurality of three- dimensional standing images; providing as input data to the machine learning model the plurality of supine spinal scans; and providing as ground truth data for the machine learning model the plurality of three- dimensional standing images.

13. The method (300) (200) of claim 12, further comprising: determining, for each of the plurality of three-dimensional standing images, a plurality of weight-bearing measurements; and including in the ground truth data the pluralities of weight-bearing measurements; wherein the plurality of weight-bearing measurements for each of the plurality of three- dimensional standing images includes at least one selected from the group consisting of an anterior disc height, a posterior disc height, a disc angle, a spondylolisthesis grade, and a spinal end plate orientation.

14. The method (300) (200) of any of claims 12 and 13, wherein generating the plurality of three-dimensional standing images includes: for each of a plurality of spinal X-Ray images, segmenting a spine in the image into a plurality of vertebra;for each of the plurality of vertebra, performing a three-dimensional pose reconstruction to generate a plurality of vertebra poses; and stitching each vertebra of the plurality of vertebra according to its predicted pose, using a lowest vertebra of the plurality of vertebra as an anchor.

15. The method (300) (200) of any of claims 12-14, further comprising: splitting the input data and the ground truth data into a plurality of training data subsets; and performing a k-fold cross validation based on the plurality of training data subsets.

Images