Therapy parameter estimation

Abstract

Aspects of the disclosed technology provide a solution to simulate a patient's treatment outcome, particularly a solution for estimating treatment parameters that can be optimally modified to achieve a desired treatment outcome. In some aspects, a process of the disclosed technology includes the steps of receiving patient data, receiving a plurality of treatment parameters, and generating an output state of the patient based on an initial state of the patient and a treatment goal of the patient. In some aspects, the process may further include calculating an objective function based on an output state of the patient and a treatment objective of the patient, and updating one or more of the treatment parameters based on the objective function. Systems and machine-readable media are also provided.

Description

Technical Field

[0001] The present technology relates to patient treatment modeling, and in particular to using treatment simulation models to predict treatment outcomes. In addition, the disclosed technology encompasses solutions for modeling how changes in specific treatment parameters may affect the ability of a given treatment plan to achieve patient goals. Background Art

[0002] A common goal of clinical interventions is to modify various structures of the patient's body, for example to achieve improved performance and / or aesthetic appearance. In this case, the goal of the clinician (e.g., physician) is to take the patient from their current condition (initial state / condition) to an end condition (treatment outcome / goal). Depending on the type of procedure, there may be many ways to achieve the goal, for example, through different implementations of the treatment plan. However, as potential treatment plans become increasingly complex, it is difficult for companies to adequately address communications with product providers. BRIEF DESCRIPTION OF THE DRAWINGS

[0003] In order to describe the manner in which the above and other advantages and features of the present disclosure can be obtained, a more particular description of the principles briefly described above will be presented by reference to specific embodiments thereof illustrated in the accompanying drawings. It should be understood that these drawings depict only exemplary embodiments of the present disclosure and are therefore not to be considered limiting of its scope. Through the use of the accompanying drawings, additional specificity and detail are used to describe and explain the principles of the present invention, in which:

[0004] Figure 1 A conceptual block diagram of an example differentiable simulator system configured to estimate treatment parameters based on clinical goals, according to some examples, is shown.

[0005] Figure 2 is a flow chart of an example process for performing treatment parameter estimation and adjustment according to some examples.

[0006] Figure 3 A conceptual block diagram of an example differentiable simulator system in which a machine learning (ML) model is used to perform treatment parameter initialization according to some examples is shown.

[0007] Figure 4 is a flow chart of an example process for training a machine learning (ML) model to estimate treatment parameters according to some examples.

[0008] Figure 5 An example of a CT scan of a patient's teeth is shown according to some examples.

[0009] Figure 6An example of a deep learning neural network that may be implemented to identify initial treatment parameters used by a differentiable simulator is shown according to some examples.

[0010] Figure 7 An example computing system according to some examples is shown. DETAILED DESCRIPTION

[0011] Various embodiments of the present disclosure are discussed in detail below. Although specific implementations are discussed, it should be understood that this is done for illustrative purposes only. Those skilled in the relevant art will recognize that other components and configurations may be used without departing from the spirit and scope of the present disclosure. Therefore, the following description and drawings are illustrative and should not be interpreted as limiting. Many specific details are described to provide a thorough understanding of the present disclosure. However, in some cases, well-known or conventional details are not described to avoid ambiguous descriptions. A reference to an embodiment or embodiment in the present disclosure may be a reference to the same embodiment or any embodiment; and such reference means at least one of the embodiments.

[0012] Reference to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. The phrase "in one embodiment" appearing in various places in the specification does not necessarily all refer to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. In addition, various features are described that may be exhibited by some embodiments but not by others.

[0013] The terms used in this specification generally have their ordinary meaning in the art in the context of this disclosure and in the specific context in which each term is used. Alternative language and synonyms may be used for any one or more of the terms discussed herein, and whether a term is elaborated or discussed herein should not be given special meaning. In some cases, synonyms for certain terms are provided. The description of one or more synonyms does not exclude the use of other synonyms. The use of examples anywhere in this specification including examples of any term discussed herein is illustrative only and is not intended to further limit the scope and meaning of the disclosure or any example term. Likewise, the disclosure is not limited to the various embodiments given in this specification.

[0014] Without intending to limit the scope of the present disclosure, examples of instruments, devices, methods and related results thereof according to embodiments of the present disclosure are given below. Note that, for the convenience of the reader, titles or subtitles may be used in the examples, which should never limit the scope of the present disclosure. Unless otherwise defined, the technical terms and scientific terms used herein have the meanings commonly understood by those of ordinary skill in the art to which the present disclosure belongs. In the event of a conflict, this document (including definitions) shall prevail.

[0015] Additional features and advantages of the present disclosure will be set forth in the following description and, in part, will be apparent from the description or can be learned by practicing the principles disclosed herein. The features and advantages of the present disclosure can be realized and obtained by means of the instruments and combinations particularly pointed out in the appended claims. These and other features of the present disclosure will become more fully apparent from the following description and the appended claims or can be learned by practicing the principles set forth herein.

[0016] Overview

[0017] In some aspects, a computer-implemented method of the disclosed technology may include: receiving patient data, the patient data including information representing an initial state of a patient; receiving multiple treatment parameters, wherein the multiple treatment parameters are based on a treatment goal of the patient; generating an output state of the patient based on the initial state of the patient and the treatment goal of the patient; calculating a loss function (or objective function) based on the output state of the patient and the treatment goal of the patient; and updating one or more of the treatment parameters based on the objective function.

[0018] A system of the disclosed technology may include: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor being configured to: receive patient data, wherein the patient data includes information representing an initial state of the patient; receive multiple treatment parameters, wherein the multiple treatment parameters are based on the treatment goals of the patient; generate an output state of the patient based on the initial state of the patient and the treatment goals of the patient; calculate an objective function (or loss function) based on the output state of the patient and the treatment goals of the patient; and update one or more of the treatment parameters based on the objective function.

[0019] A non-transitory computer-readable storage medium may include at least one instruction for causing a computer or processor to: receive patient data, wherein the patient data includes information representing an initial state of a patient; receive multiple treatment parameters, wherein the multiple treatment parameters are based on a treatment goal of the patient; generate an output state of the patient based on the initial state of the patient and the treatment goal of the patient; calculate an objective function based on the output state of the patient and the treatment goal of the patient; and update one or more of the treatment parameters based on the objective function.

[0020] In some aspects, a computer-implemented method of the disclosed technology may include the following steps: receiving patient data associated with a patient through a machine learning (ML) model; receiving a treatment goal associated with the patient through the ML model; predicting one or more estimated treatment parameters based on the patient data and the treatment goal through the ML model; providing the one or more estimated treatment parameters to a simulation model; generating an output state for the patient based on the one or more estimated treatment parameters through the simulation model; and calculating an objective function based on the output state of the patient and the treatment goal of the patient.

[0021] A system of the disclosed technology may include: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor configured to: receive patient data associated with a patient through a machine learning (ML) model; receive a treatment goal associated with the patient through the ML model; predict one or more estimated treatment parameters based on the patient data and the treatment goal through the ML model; provide the one or more estimated treatment parameters to a simulation model; generate an output state of the patient based on the one or more estimated treatment parameters through the simulation model; and calculate an objective function based on the output state of the patient and the treatment goal of the patient.

[0022] A non-transitory computer-readable storage medium (e.g., a program product) of the disclosed technology may include at least one instruction for causing a computer or processor to: receive patient data associated with a patient through a machine learning (ML) model; receive a treatment goal associated with the patient through the ML model; predict one or more estimated treatment parameters based on the patient data and the treatment goal through the ML model; provide the one or more estimated treatment parameters to a simulation model; generate an output state for the patient based on the one or more estimated treatment parameters through the simulation model; and calculate an objective function based on the output state of the patient and the treatment goal of the patient.

[0023] describe

[0024] As previously mentioned, a patient treatment plan can be used to achieve a desired treatment outcome or patient treatment goal. According to the treatment plan, different treatment parameters can be applied and modified during the treatment process to affect the patient's results. Treatment parameters can include almost any condition, variable and / or procedure that can affect the treatment outcome. For example, treatment parameters can include material parameters, including the composition, size and / or shape of the instrument used to implement the treatment. In some methods, treatment parameters can identify specific parts or regions (landmarks) of the patient's anatomical structure, which will be changed as a target by the applied treatment. Treatment parameters can also describe treatment techniques or procedures, such as the timing and / or sequence of the applied procedures. For example, an orthodontic procedure can involve repositioning the patient's teeth to a desired arrangement in order to correct malocclusion and / or improve aesthetics. To achieve these purposes, orthodontic appliances such as braces, shell appliances, etc. can be applied to the patient's teeth by an orthodontist. The appliance can be configured to apply force on one or more teeth in order to achieve the desired tooth movement according to the treatment plan. Treatment parameters in this context may include the type of appliance used (e.g., braces or shell appliances), the material properties of the appliance selected (which affect the forces applied to each tooth), and / or the treatment regimen (e.g., when the appliance is introduced into treatment, and / or the duration of use, etc.) Therefore, treatment parameters are critical to achieving the desired treatment goals.

[0025] With the increasing use of three-dimensional (3D) computer graphics and software in clinical applications, it would be useful to use computational methods to model the effects of various treatment parameters on clinical outcomes. In a similar manner, it would be useful to model treatment outcomes in reverse order, for example to determine which treatment parameters can be implemented to achieve a desired goal. However, due to the large number of potential treatment parameters associated with even routine procedures, modeling methods (e.g., those utilizing naive optimization strategies) are computationally intractable in most real-world scenarios.

[0026] Various aspects of the disclosed technology provide solutions for modeling treatment outcomes, and specifically provide solutions for identifying and modifying specific treatment parameters to achieve clinical goals. Treatment outcomes can be predicted (modeled) using a differentiable simulator that is configured to receive input data representing the current patient state and treatment plan parameters and model (one or more) corresponding treatment outcomes. In some aspects, patient data can represent physical patient attributes. For example, patient data can include a biomechanical model representing the geometry of a particular anatomical structure. In the orthodontic context, such a model can include, but is not limited to, image data (e.g., photographs), cone beam computed tomography (CBCT) scan data, x-ray images, magnetic resonance imaging (MRI) data, intraoral scans, and / or facial scans, etc.

[0027] The modeled patient outcomes can be used to calculate an error function (or objective function) that represents the difference between the desired goal and the modeled patient outcomes. Thus, in some methods, optimal patient treatment methods (e.g., optimal combination of treatment parameters) are those methods that minimize the deviation between the modeled patient outcomes and the treatment goals, such as by minimizing the error function. As discussed in further detail below, differentiable simulation methods (e.g., using the first-order derivative of the error function) can be used to identify treatment parameters that can be adjusted / updated to most effectively minimize the error function.

[0028] In some embodiments, machine learning methods can be used to identify / initialize treatment parameters, such as for use in a differentiable simulator. In such methods, a machine learning (ML) model trained to identify target treatment parameters (e.g., based on an expected goal) can be used to select / initialize treatment parameters for a differentiable simulator. As discussed in further detail below, parameter update techniques that further adjust parameters based on derivatives (e.g., first-order derivatives or second-order derivatives, etc.) of an error function (or objective function) can be used to further adjust / update the ML-selected treatment parameters.

[0029] Several of the examples provided relate to clinical interventions in the context of orthodontic treatment, however it should be understood that the concepts disclosed herein can be applied to other clinical applications without departing from the scope of the disclosed technology. For example, aspects of the disclosed technology can be applicable to a variety of patient treatments, including surgical interventions for modifying various aspects of a patient's anatomy. For example, the disclosed methods can be used in the context of orthognathic surgical procedures designed to change a patient's jaw and / or facial shape, etc.

[0030] Figure 1A conceptual block diagram of an example differentiable simulator system 100 configured to estimate treatment parameters based on clinical goals according to some examples is shown. The simulator system 100 includes a treatment plan 102, which includes treatment parameters 103 and treatment goals 105. As described above, the treatment parameters 103 may include any metric or variable that can describe any aspect or characteristic of a clinical intervention administered to a patient. The treatment goals 105 may include data describing an expected (or target) patient outcome. By way of example with respect to the context of orthodontic treatment, the treatment parameters 103 may include parameters describing the material composition and / or shape of an aligner tray (or invisible braces) configured to reposition the patient's teeth in a predetermined manner. The treatment parameters 103 may also describe the forces acting on different portions or locations of the patient's anatomy due to the characteristics of the bracket construction, such as forces applied to individual teeth or groups of teeth. The treatment goals 105 may in turn include data indicating a target dentition for the patient. For example, the treatment goals 105 may include a photograph and / or a three-dimensional (3D) mesh (or other 3D model representation) of an expected or desired arrangement of the patient's teeth and / or a desired aesthetic outcome.

[0031] In practice, the treatment parameters 103 are provided to the simulator 108 along with the patient data 106. The patient data 106 received by the simulator 108 may include any data representing the current or most recent condition of the patient. For example, the patient data may include measurements or images representing portions of the patient's anatomy that will be altered by the therapeutic intervention. In some aspects, the patient data 106 may be represented as a biomechanical model of the patient's anatomy. For example, in an orthodontic context, the patient data may include images of teeth and / or facial features, including but not limited to digital images, intraoral scans, cone beam computed tomography (CBCT) scans, x-ray images, and the like.

[0032] Using the received treatment parameters 103 and patient data 106, the simulator 108 can model the expected patient outcome (or output state), for example, as represented by output data 110. In some aspects, the simulator 108 can be (or can utilize) a physics-based model, such as a biomechanical model, which is configured to predict the outcome of the treatment parameters on the provided patient state indicated by the patient data 106. For example, the simulator 108 can be configured to utilize numerical methods (e.g., such as the Finite Element Method (FEM), etc.) to model the structural mechanics of the patient's anatomical structure when the patient's anatomical structure is treated using the intervention reflected by the treatment parameters 103. Numerical modeling can be used to simulate the behavior of the patient's anatomical structure when the patient's anatomical structure is treated using the intervention reflected by the treatment parameters 103. Examples of numerical methods that can be used in such modeling include the Finite Element Method (FEM), the Boundary Element Method (BEM), the Finite Difference Method (FDM), and the Discrete Element Method (DEM). Embodiments are discussed with reference to the finite element method (FEM), finite element models, and applications of finite element analysis. However, it should be understood that any other possible type of numerical analysis and modeling may be performed in embodiments instead of FEM. Using an orthodontic example, the simulator 108 may be configured to model (predict) the movement of a patient's teeth in response to the application of an orthodontic liner. In this example, the treatment parameters 103 may include data reflecting the size, shape, and / or material properties of the liner, while the patient data 106 may include data reflecting the current position and orientation of the patient's teeth and other oral structures, such as an image or a 3D mesh model.

[0033] The predicted impact of the applied treatment parameters 103 is reflected in the output data 110 (or output state) generated by the simulator 108. For example, depending on the desired implementation, the output data 110 can be (or can include) simulated post-treatment data in the form of a 3D mesh, or image data, which reflects the structural and / or aesthetic changes of the patient after the intervention using the treatment parameters 103. The output data 110 can then be compared with the treatment goals 105 to determine / calculate an objective function 112 that represents the similarity between the post-treatment results (e.g., as the output data 110) and the patient's clinical goals (e.g., goals 105).

[0034] In some aspects, the objective function 112 can be used to identify (e.g., from all of the treatment parameters 103) specific treatment parameters that most significantly affect the post-treatment outcome. For example, the first-order derivative of the objective function 112 can be used to identify one or more treatment parameters that contribute most to the reduction in magnitude associated with the objective function 112. The treatment parameters identified from the first-order derivative of the objective function (or loss function) 112 can be selected for modification / updating, for example, using a parameter update operation 114. In this case, the previously applied treatment parameters 103 used by the simulator 108 can be modified based on the parameter selection performed using the objective function 112. The updated or adjusted parameters can then be propagated back into the treatment plan 102, for example, as an update to the treatment parameters 103 used by the simulator 108. The process of forward modeling the patient outcome (e.g., as reflected in the output data 110) and adjusting the treatment parameters 103 used by the simulator 108 can be iterated until a satisfactory patient outcome is achieved, i.e., until the treatment goal 105 is achieved.

[0035] For example, in the context of an orthodontic intervention, the simulator 108 can be used to predict the resulting geometry of the patient's teeth (e.g., as reflected in the output data 110) after an intervention utilizing the treatment parameters (103) and based on the patient's current dentition (e.g., as reflected in the patient data 106). The output data 110 can then be used to determine (or calculate) an objective function 112, to which the differential analysis can be applied. For example, the differential analysis of the resulting objective function 112 can indicate that certain treatment parameters (such as appliance composition) may be more meaningful for achieving the treatment goal 105 than other / different treatment parameters (such as appliance geometry). Additionally, for each treatment parameter, the derivative can indicate the magnitude and direction of the update, e.g., to minimize loss (error). Thus, by utilizing the differential analysis of the objective (error) function 112, the treatment parameters 103 can be more effectively adjusted / updated, thereby improving patient treatment outcomes.

[0036] Figure 2 is a flow chart of an example process 200 for performing treatment parameter estimation and adjustment according to some examples. At step 202, a differentiable simulator (such as the one described above with reference to Figure 1The simulator 108 discussed above receives patient data. The received patient data may represent an initial (or current) state of the patient, such as information describing a portion of the patient's anatomical structure. As described above, the patient data may include a biomechanical model representing the geometry of a particular anatomical structure. In the orthodontic context, such a model may include, but is not limited to, image data (e.g., photographs), cone beam computed tomography (CBCT) scan data, x-ray images, magnetic resonance imaging (MRI) data, intraoral scans, and / or facial scans, etc.

[0037] At step 204, a plurality of treatment parameters may be received, for example, by a differentiable simulator. As described above, treatment parameters may include material parameters, including the composition, size and / or shape of the apparatus used to implement the treatment. Treatment parameters may also describe treatment techniques or procedures, such as the timing and / or sequence of the procedures applied, etc. For example, an orthodontic procedure may involve repositioning the patient's teeth to a desired arrangement in order to correct malocclusion and / or improve aesthetics. To achieve these purposes, orthodontic appliances such as braces, shell appliances, etc. may be applied to the patient's teeth by an orthodontist. The appliance may be configured to apply forces on one or more teeth in order to achieve the desired tooth movement according to a treatment plan. Treatment parameters in this context may include the type of appliance used (e.g., braces or shell appliances), the material properties of the selected appliance (which affect the forces applied to each tooth), and / or a treatment regimen (such as when the appliance is introduced into treatment, and / or the duration of use, etc.).

[0038] At step 206, process 200 includes generating an output state for the patient based on the initial state of the patient (e.g., based on the patient data) and the received treatment parameters. Figure 1 As discussed, the output state may represent, for example, a simulation of a post-treatment result resulting from the application of the treatment parameters. Thus, the output state (or output data) may be (or may include), for example, simulated post-treatment data in the form of a 3D mesh, or image data, which reflects the structural and / or aesthetic changes of the patient after the intervention using the treatment parameters.

[0039] As described above, the output state can be an output generated by a simulation model, such as a differentiable simulator, which is (or includes) a physics-based model. For example, the simulator (or differentiable simulator) can be configured to utilize a finite element method (FEM) to model the structural mechanics of the patient's anatomical structure when the patient's anatomical structure is treated using an intervention reflected by the received treatment parameters.

[0040] At step 208, process 200 includes calculating (or determining) an objective function based on the patient's output state and treatment goals. The objective function can represent the degree of similarity between the post-treatment results (e.g., as indicated by the output state / data) and the patient's clinical goals. As described above, a loss function can be used to identify specific treatment parameters that most significantly affect the treatment results. For example, a first-order derivative of the objective function can be used to identify treatment parameters for modification / updating, such as to better achieve the patient's treatment goals.

[0041] At step 210, process 200 includes updating one or more of the treatment parameters based on the objective function. In such an approach, certain (selected) previously applied treatment parameters may be modified based on the determined contribution to the objective function. For example, treatment parameters that contribute more meaningfully to the overall loss (or error) may be preferentially selected for modification / updating, i.e., to improve the outcome of the administered treatment. The updated / adjusted treatment parameters may then be propagated as input to the simulator.

[0042] In some aspects, machine learning (ML) methods can be used to initialize treatment plan parameters. For example, treatment parameters can be provided as output by an ML model that is trained to estimate / predict optimal treatment parameters given paired treatment goal and patient data inputs. Further details on using ML methods to perform parameter estimation / initialization are provided below. Figure 3 and Figure 4 supply.

[0043] Figure 3 A conceptual block diagram of an example differentiable simulator system 300 is shown in which a machine learning (ML) model 302 is used to perform treatment parameter initialization according to some examples. Figure 3 In the example of , ML model 302 is configured to predict / estimate treatment parameters 303, for example, based on paired inputs of treatment goal 305 and patient data 306. Figure 2In the example discussed, the treatment goal may include any information representing a desired or expected patient outcome. For example, the treatment goal may include a representation of the target post-treatment outcome, including, but not limited to, image and / or 3D mesh data, etc. In addition, the treatment goal data 305 may include other types of imaging data, including, but not limited to, cone beam computed tomography (CBCT) scan data, x-ray images, magnetic resonance imaging (MRI) data, intraoral scans, and / or facial scans, etc. Similarly, the patient data 306 may include data representing the patient's current (or most recent) condition (such as, one or more areas of the patient's anatomy that will be changed by the treatment plan). Therefore, the patient data may also include image data (e.g., photographs), cone beam computed tomography (CBCT) scan data, x-ray images, magnetic resonance imaging (MRI) data, intraoral scans, and / or facial scans, etc.

[0044] Based on the received patient target data 305 and patient data 306, the ML model can predict / estimate initial treatment parameters 303, which are then provided to a simulator (e.g., a differentiable simulator 308). Figure 1 As discussed above, the simulator 308 may use the received treatment parameters to model the effects of the applied treatment plan and output a predicted post-treatment outcome as output data 310. Depending on the desired implementation, the simulator 308 may be (or may utilize) a physics-based model (such as a biomechanical model) configured to predict the outcome of the treatment parameters on a provided patient state indicated by the patient data 106.

[0045] In some aspects, the output data 310 can be used to evaluate the simulated post-treatment outcomes based on similarity to the treatment objectives (i.e., as defined by the treatment goals 305). As described above, this evaluation can be performed using a target (error) function 312 that defines the difference between the target treatment goals 305 and the post-treatment patient outcomes reflected in the output data 310. In some embodiments, the target error function 312 can be used to identify and update / adjust selected parameters, as described above with respect to Figure 1 and Figure 2 For example, using differential analysis (e.g., by calculating the first derivative of the objective error function 312), the treatment parameters that have the greatest impact on the output data 310 can be identified. By updating / adjusting these selected parameters (e.g., in a parameter update step 314), the input parameters 303 of the simulator 308 can be modified to better achieve the desired clinical outcome.

[0046] Figure 44 is a flow chart of an example process 400 for training a machine learning (ML) model to estimate treatment parameters according to some examples. The process 400 begins with step 402, in which patient data is received. The patient data can represent physical patient attributes. For example, the patient data can include a biomechanical model representing a particular anatomical structure of the patient. For example, the patient data can include, but is not limited to, image data (e.g., photographs), cone beam computed tomography (CBCT) scan data, x-ray images, magnetic resonance imaging (MRI) data, intraoral scans, and / or facial scans, etc.

[0047] At step 404, process 400 includes receiving treatment goals associated with the patient. As described above, the treatment goals may include data specifying the desired / target outcome of the treatment plan. Similar to the patient data, the treatment goal data may include, but is not limited to, image data (e.g., photographs), cone beam computed tomography (CBCT) scan data, x-ray images, magnetic resonance imaging (MRI) data, intraoral scans, and / or facial scans, etc.

[0048] At step 406, process 400 may include predicting one or more (estimated) treatment parameters using the ML model, for example based on the patient data and the treatment goals. As described above, the ML model may be configured to predict optimal treatment parameters to achieve desired patient goals.

[0049] At step 408, process 400 may include providing the estimated treatment parameters to a simulation model, e.g., the simulation model configured to simulate post-treatment data of the patient based on the applied treatment parameters. In some aspects, the simulation model may be (or may utilize) a physics-based model (such as a biomechanical model) configured to predict the results of the treatment parameters on a provided patient state (as indicated by the patient data received at step 402). For example, the simulation model may be configured to utilize a finite element method (FEM) to model the structural mechanics of the patient's anatomical structure as the patient's anatomical structure is treated using the intervention reflected by the treatment parameters 103.

[0050] Using the orthodontic example, the simulation model can be configured to model (predict) the movement of the patient's teeth in response to the application of an orthodontic liner. In this example, the treatment parameters may include the size, shape, and / or material properties of the liner, while the patient data may include image data and / or a 3D mesh model reflecting the current position and orientation of the patient's teeth and / or soft tissue structures.

[0051] In step 410, process 400 includes generating an output state for the patient based on the initial state of the patient (e.g., based on the patient data) and the received treatment parameters. As described above, the output state can represent, for example, a simulation of a post-treatment result resulting from the application of the treatment parameters. Thus, the output state (or output data) can include, for example, simulated post-treatment data in the form of a 3D mesh, or image data that reflects structural and / or aesthetic changes in the patient after the intervention using the treatment parameters.

[0052] At step 412, process 400 includes calculating an objective function (also: loss function) based on the patient's output state and the patient's treatment goal. The objective function can be used to determine the success (or failure) of the applied treatment, for example, by representing the difference between the patient's state after treatment and the treatment goal. In some aspects, the target loss can be used to update one or more weights of the machine learning model in order to perform training. In some aspects, the derivative of the objective function (loss function) can be used to determine the magnitude and direction (sign) of the ML model weights to be updated. Through training, the calculated objective function can be used to improve the ML model predictions about future treatment parameters. In some aspects, once training is completed and put into use, the loss function can be used to further update the treatment parameters provided to the simulation model, for example, where additional simulation iterations are required. For example, differential analysis techniques can be used to identify / select specific treatment parameters for updating / adjusting, which can then be used to perform additional treatment simulations.

[0053] Figure 5 An exemplary diagram 500 of a CT scan of a tooth is shown. In this embodiment, the tooth roots are directly derived from a high-resolution CBCT scan of the patient. The scanned roots can then be applied to crowns derived from impressions, or used with existing crowns extracted from cone beam computed tomography (CBCT) data. A single CBCT scan gives 3D data and multiple forms of X-ray-like data. PVS impressions are avoided.

[0054] In one embodiment, a cone-beam x-ray source and a 2D area detector scan the patient's dental anatomy by any of a variety of methods, preferably over a 360 degree angular range and along its entire length, wherein the position of the area detector is fixed relative to the source, and relative rotational and translational movement between the source and the object provides scanning (irradiation of the object with radiant energy). As a result of relative movement of the cone-beam source along a scan path to a plurality of source positions (i.e., "views"), the detector acquires corresponding multiple sets of sequential cone-beam projection data (also referred to herein as cone-beam data or projection data), each set of cone-beam data representing x-ray attenuation caused by the object at a corresponding one of the source positions.

[0055] The present disclosure now turns to further discussion of models that can be used by the environment and techniques described herein. Specifically, Figure 6 is an illustrative example of a deep learning neural network 600 that can be implemented to perform treatment parameter estimation. As described above, the input layer 620 can be configured to receive patient data and / or data related to a target treatment goal. The neural network 600 includes a plurality of hidden layers 622a, 622b to 622n. The hidden layers 622a, 622b to 622n include "n" hidden layers, where "n" is an integer greater than or equal to 1. The number of hidden layers can be made to include as many layers as required for a given application. The neural network 600 also includes an output layer 621, which provides an output generated by the processing performed by the hidden layers 622a, 622b to 622n. In an illustrative example, the output layer 621 can provide an estimated treatment parameter (e.g., the estimated parameter 303), which can be used / received by a differential simulator to estimate a patient treatment outcome.

[0056] Neural network 600 is a multi-layer neural network of interconnected nodes. Each node can represent a piece of information. Information associated with a node is shared between different layers, and each layer retains the information as it is processed. In some cases, neural network 600 may include a feedforward network, in which case there is no feedback connection in which the output of the network is fed back into itself. In some cases, neural network 600 may include a recursive neural network, which may have loops that allow information to be carried across nodes when reading in input.

[0057] Information can be exchanged between nodes by node-to-node interconnection between each layer. The nodes of input layer 620 can activate the node group in the first hidden layer 622a. For example, as shown in the figure, each of the input nodes of input layer 620 is connected to each of the nodes of the first hidden layer 622a. The nodes of the first hidden layer 622a can transform the information of each input node by applying an activation function to the input node information. Then, the information derived from the transformation can be passed to the nodes of the next hidden layer 622b and can activate the nodes of the next hidden layer 622b, and the nodes of the next hidden layer 622b can perform their own specified functions. Example functions include convolution function, upsampling function, data transformation function and / or any other suitable function. Then, the output of hidden layer 622b can activate the nodes of the next hidden layer, and so on. The output of the last hidden layer 622n can activate one or more nodes of output layer 621, and output is provided at one or more nodes of the output layer 621. In some cases, although a node in neural network 600 (e.g., node 626) is shown as having multiple output lines, the node may have a single output, and all lines shown as output from the node represent the same output value.

[0058] In some cases, each node or interconnection between nodes can have a weight that is a set of parameters derived from the training of neural network 600. Once neural network 600 is trained, it can be referred to as a trained neural network, which can be used to classify one or more activities. For example, an interconnection between nodes can represent a piece of information learned about the interconnected nodes. The interconnections can have adjustable digital weights that can be adjusted (e.g., based on a training data set), allowing neural network 600 to adapt to the input and learn as more data is processed.

[0059] The neural network 600 is pre-trained to process features of data from the input layer 620 using different hidden layers 622a, 622b to 622n to provide an output through the output layer 621.

[0060] In some cases, neural network 600 can use a training process called back propagation to adjust the weights of the nodes. As described above, the back propagation process can include a forward pass, a loss function, a backward pass, and a weight update. The forward pass, loss function, backward pass, and parameter update are performed for one training iteration. This process can be repeated for a certain number of iterations for each training data set until neural network 600 is trained well enough so that the weights of the layers are accurately adjusted.

[0061] As described above, for the first training iteration of the neural network 600, the output will likely include values ​​that do not prioritize any particular class due to the weights being randomly selected at initialization. For example, if the output is a vector with probabilities that the object includes different classes, the probability values ​​for each of the different classes may be equal or at least very similar (e.g., for ten possible classes, each class may have a probability value of 0.1). Using the initial weights, the neural network 600 cannot determine low-level features and therefore cannot accurately determine what the classification of the object may be. A loss function can be used to analyze the errors in the output. Any suitable loss function definition can be used, such as a cross entropy loss. Another example of a loss function includes the mean squared error (MSE), which is defined as The loss can be set equal to the value of E_total.

[0062] For the first training image, the loss (or error) will be high because the actual value will be very different from the predicted output. The goal of training is to minimize the amount of loss so that the predicted output is the same as the training label. The neural network 600 can perform a back pass by determining which inputs (weights) contribute the most to the loss of the network, and the weights can be adjusted so that the loss is reduced and ultimately minimized. The derivative of the loss with respect to the weight (expressed as dL / dW, where W is the weight at a particular layer) can be calculated to determine the weight that contributes the most to the loss of the network. After the derivative is calculated, the weight update can be performed by updating all the weights of the filter. For example, the weights can be updated so that they change in the opposite direction of the gradient. The weight update can be expressed as w=w_i-ηdL / dW, where w represents the weight, wi represents the initial weight, and η represents the learning rate. The learning rate can be set to any suitable value, where a high learning rate includes a larger weight update, while a lower value indicates a smaller weight update.

[0063] The neural network 600 may include any suitable deep network. One example includes a convolutional neural network (CNN), which includes an input layer and an output layer with multiple hidden layers between the input layer and the output layer. The hidden layers of the CNN include a series of convolutional layers, non-linear layers, pooling layers (for downsampling), and fully connected layers. The neural network 600 may include any other deep network other than a CNN, such as an autoencoder, a deep belief network (DBN), a recursive neural network (RNN), etc.

[0064] As will be appreciated by those skilled in the art, machine learning-based classification techniques may vary depending on the desired implementation. For example, machine learning classification schemes may utilize one or more of the following alone or in combination: Hidden Markov Models; Recursive Neural Networks; Convolutional Neural Networks (CNNs); Deep Learning; Bayesian Symbolic Methods; Generative Adversarial Networks (GANs); Support Vector Machines; Image Registration Methods; Applicable Rule-Based Systems. In the case of using regression algorithms, they may include, but are not limited to, Stochastic Gradient Descent Regressors and / or Passive Aggressive Regressors, etc.

[0065] The machine learning classification model can also be based on a clustering algorithm (e.g., a mini-batch K-means clustering algorithm), a recommendation algorithm (e.g., a Miniwise Hashing algorithm or an Euclidean Locality-Sensitive Hashing (LSH) algorithm), and / or an anomaly detection algorithm (such as a local outlier factor). In addition, the machine learning model can adopt a dimensionality reduction method, such as one or more of the following: a mini-batch dictionary learning algorithm, an incremental principal component analysis (PCA) algorithm, a latent Dirichlet Allocation algorithm, and / or a mini-batch K-means algorithm.

[0066] This disclosure now turns to Figure 7, which shows an example of a processor-based computing system 700 in which the components of the system communicate with each other electrically using a bus 705. The computing system 700 may include a processing unit (CPU or processor) 710 and a system bus 705, which may couple various system components including a system memory 715 (such as a read-only memory (ROM) 720 and a random access memory (RAM) 725) to the processor 710. The computing system 700 may include a cache 712 of high-speed memory directly connected to the processor 710, adjacent to the processor 710, or integrated as part of the processor 710. The computing system 700 may copy data from the memory 715, ROM 720, RAM 725, and / or storage device 730 to the cache 712 for quick access by the processor 710. In this way, the cache 712 may provide a performance boost that avoids processor delays while waiting for data. These and other modules may control the processor 710 to perform various actions. Other system memories 715 may also be available. The memory 715 may include a variety of different types of memory with different performance characteristics. The processor 710 may include any general purpose processor and hardware modules or software modules configured to control the processor 710 (such as module 1 732, module 2 734, and module 3 736 stored in the storage device 730) and a dedicated processor in which the software instructions are incorporated into the actual processor design. The processor 710 may essentially be a completely self-contained computing system, including multiple cores or processors, buses, memory controllers, caches, etc. Multi-core processors may be symmetric or asymmetric.

[0067] In order to enable user interaction with the computing system 700, the input device 745 can represent any number of input mechanisms, such as a microphone for voice, a touch-protected screen for gesture or graphic input, a keyboard, a mouse, motion input, voice, etc. The output device 735 can also be one or more of a plurality of output mechanisms known to those skilled in the art. In some cases, a multimodal system can enable a user to provide multiple types of input to communicate with the computing system 700. The communication interface 740 can govern and manage user input and system output. There may be no restrictions on operating on any particular hardware arrangement, and therefore the basic features here can be easily replaced by improved hardware or firmware arrangements as they are developed.

[0068] The storage device 730 may be a non-volatile memory and may be a hard disk or other type of computer-readable medium that can store data that can be accessed by a computer, such as a magnetic tape cartridge, a flash memory card, a solid-state memory device, a digital versatile disk, a cassette, a random access memory, a read-only memory, and a mixture thereof.

[0069] As described above, the storage device 730 may include software modules 732, 734, 736 for controlling the processor 710. Other hardware or software modules are contemplated. The storage device 730 may be connected to the system bus 705. In some embodiments, a hardware module that performs a particular function may include a software component stored in a computer-readable medium that is connected to the necessary hardware components (such as the processor 710, the bus 705, the output device 735, etc.) to perform the function. In some cases, for clarity of explanation, the present technology may be presented as including separate functional blocks, including functional blocks that include devices, device components, steps in methods embodied in software or a combination of hardware and software, or routines.

[0070] In some embodiments, computer-readable storage devices, media, and memories may include cables or wireless signals containing bit streams, etc. However, when referred to, non-transitory computer-readable storage media explicitly excludes media such as energy, carrier signals, electromagnetic waves, and signals themselves.

[0071] The method according to the above example can be implemented using computer executable instructions stored in a computer readable medium or otherwise obtained from a computer readable medium. Such instructions may include, for example, instructions and data that cause or otherwise configure a general-purpose computer, a special-purpose computer, or a special-purpose processing device to perform a specific function or functional group. The part of the computer resources used can be accessed through a network. Computer executable instructions can be, for example, binary files, intermediate format instructions (such as assembly language), firmware, or source code. Examples of computer readable media that can be used to store instructions, information used, and / or information created during the method according to the described examples include disks or optical disks, flash memory, USB devices provided with non-volatile memory, networked storage devices, etc.

[0072] Devices implementing methods according to these disclosures may include hardware, firmware, and / or software, and may be implemented in any of a variety of form factors. Typical examples of such form factors include laptop computers, smart phones, small form factor personal computers, personal digital assistants, rack-mounted devices, stand-alone devices, etc. The functions described herein may also be embodied in peripheral devices or add-on cards. As another example, such functions may also be implemented on a circuit board between different chips or in different processes executed in a single device.

[0073] Instructions, the medium for transmitting such instructions, computing resources for executing them, and other structures for supporting such computing resources are the means for providing the functions described in these disclosures.

[0074] Although various examples and other information are used to explain aspects within the scope of the appended claims, limitations on the claims should not be implied based on specific features or arrangements in such examples, because a person of ordinary skill will be able to use these examples to obtain a variety of implementations. In addition, although some topics may have been described in language specific to examples of structural features and / or method steps, it should be understood that the subject matter defined in the appended claims is not necessarily limited to these described features or actions. For example, such functions may be distributed differently in components other than those identified herein or performed in components other than those identified herein. Instead, the described features and steps are disclosed as examples of components of systems and methods within the scope of the appended claims.

[0075] Claim language reciting "at least one of..." refers to at least one of a set and indicates that one member of the set or multiple members of the set satisfy the claim. For example, claim language reciting "at least one of A and B" means A, B, or A and B.

Claims

1. A computer-implemented method comprising: receiving patient data, wherein the patient data includes information indicating an initial state of the patient; receiving a plurality of treatment parameters, wherein the plurality of treatment parameters are based on treatment goals for the patient; generating an output state for the patient based on the initial state of the patient and the plurality of treatment parameters; calculating an objective function based on the patient's output state and the patient's treatment goal; and One or more of the treatment parameters are updated based on the objective function.

2. The computer-implemented method of claim 1 , wherein: Updating the one or more treatment parameters further comprises: The one or more treatment parameters are identified based on a derivative of the objective function.

3. The computer-implemented method of claim 1 , wherein: Generating the output status of the patient further comprises: The patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a numerical model.

4. The computer-implemented method of claim 3, wherein: The numerical model includes a biomechanical model.

5. The computer-implemented method of claim 1 , wherein: The output state of the patient comprises a three-dimensional 3D mesh.

6. The computer-implemented method of claim 1 , wherein: The patient data includes an intraoral scan.

7. The computer-implemented method of claim 1 , wherein: The patient's treatment target comprises a three-dimensional 3D mesh.

8. A system comprising: at least one memory; as well as at least one processor, coupled to the at least one memory, the at least one processor being configured to: receiving patient data, wherein the patient data includes information indicating an initial state of the patient; receiving a plurality of treatment parameters, wherein the plurality of treatment parameters are based on treatment goals for the patient; generating an output state for the patient based on the initial state of the patient and the plurality of treatment parameters; calculating an objective function based on the patient's output state and the patient's treatment goal; and One or more of the treatment parameters are updated based on the objective function.

9. The system according to claim 8, wherein: To update the one or more treatment parameters, the at least one processor is further configured to: The one or more treatment parameters are identified based on a derivative of the objective function.

10. The system according to claim 8, wherein: To generate an output state of the patient, the at least one processor is further configured to: The patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a numerical model.

11. The system according to claim 10, wherein: The numerical model includes a biomechanical model.

12. The system according to claim 8, wherein: To generate an output state of the patient, the at least one processor is further configured to: The patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a machine learning (ML) model.

13. The system according to claim 8, wherein: The patient data includes an intraoral scan.

14. The system according to claim 8, wherein: The patient's treatment goals include a three-dimensional grid.

15. A non-transitory computer-readable storage medium comprising at least one instruction for causing a computer or a processor to: Receiving patient data, wherein The patient data includes information representing an initial state of the patient; receiving a plurality of treatment parameters, wherein the plurality of treatment parameters are based on treatment goals for the patient; generating an output state for the patient based on the initial state of the patient and the plurality of treatment parameters; calculating an objective function based on the patient's output state and the patient's treatment goal; and One or more of the treatment parameters are updated based on the objective function.

16. The non-transitory computer-readable storage medium of claim 15, wherein: To update the one or more treatment parameters, the at least one instruction is further configured to cause the processor to: The one or more treatment parameters are identified based on a derivative of the objective function.

17. The non-transitory computer-readable storage medium of claim 15, wherein: To generate an output state for the patient, the at least one instruction is further configured to cause the processor to: The patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a numerical model.

18. The non-transitory computer-readable storage medium of claim 17, wherein: The numerical model includes a biomechanical model.

19. The non-transitory computer-readable storage medium of claim 15, wherein: To generate an output state of the patient, the at least one processor is further configured to: The patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a machine learning (ML) model.

20. The non-transitory computer-readable storage medium of claim 15, wherein: The patient data includes an intraoral scan.

21. A computer-implemented method comprising: receiving patient data associated with the patient via a machine learning (ML) model; receiving, by the ML model, a treatment goal associated with the patient; predicting, by the ML model, one or more estimated treatment parameters based on the patient data and the treatment goal; providing the one or more estimated treatment parameters to a simulation model; generating, by the simulation model, an output state for the patient based on the one or more estimated treatment parameters; as well as An objective function is calculated based on the patient's output state and the patient's treatment goal.

22. The computer-implemented method of claim 21 , further comprising: One or more of the treatment parameters are updated based on the objective function.

23. The computer-implemented method of claim 22, wherein: Updating the one or more treatment parameters further comprises: The one or more treatment parameters are identified based on a derivative of the objective function.

24. The computer-implemented method of claim 21, wherein: The simulation model includes a numerical model.

25. The computer-implemented method of claim 24, wherein: The numerical model includes a biomechanical model.

26. The computer-implemented method of claim 21, wherein: The treatment target includes a three-dimensional 3D model.

27. The computer-implemented method of claim 21, wherein: The patient data includes an intraoral scan.

28. A system comprising: at least one memory; as well as at least one processor, coupled to the at least one memory, the at least one processor being configured to: receiving patient data associated with the patient via a machine learning (ML) model; receiving, by the ML model, a treatment goal associated with the patient; predicting, by the ML model, one or more estimated treatment parameters based on the patient data and the treatment goal; providing the one or more estimated treatment parameters to a simulation model; generating, by the simulation model, an output state for the patient based on the one or more estimated treatment parameters; as well as An objective function is calculated based on the patient's output state and the patient's treatment goal.

29. The system of claim 28, wherein: The at least one processor is further configured to: One or more of the treatment parameters are updated based on the objective function.

30. The system of claim 29, wherein: To update the one or more treatment parameters, the at least one processor is further configured to: The one or more treatment parameters are identified based on a derivative of the objective function.

31. The system of claim 28, wherein: The simulation model includes a numerical model.

32. The system of claim 31, wherein: The numerical model includes a biomechanical model.

33. The system of claim 28, wherein: The treatment target includes a three-dimensional 3D model.

34. The system of claim 28, wherein: The patient data includes an intraoral scan.

35. A non-transitory computer-readable storage medium comprising at least one instruction for causing a computer or a processor to: receiving patient data associated with the patient via a machine learning (ML) model; receiving, by the ML model, a treatment goal associated with the patient; predicting, by the ML model, one or more estimated treatment parameters based on the patient data and the treatment goal; providing the one or more estimated treatment parameters to a simulation model; generating, by the simulation model, an output state for the patient based on the one or more estimated treatment parameters; as well as An objective function is calculated based on the patient's output state and the patient's treatment goal.

36. The non-transitory computer-readable storage medium of claim 35, wherein: The at least one instruction is further configured to cause the computer or processor to: One or more of the treatment parameters are updated based on the objective function.

37. The non-transitory computer-readable storage medium of claim 36, wherein: To update the one or more treatment parameters, wherein the at least one instruction is further configured to cause the computer or processor to: The one or more treatment parameters are identified based on a derivative of the objective function.

38. The non-transitory computer-readable storage medium of claim 35, wherein: The simulation model includes a numerical model.

39. The non-transitory computer-readable storage medium of claim 38, wherein: The numerical model includes a biomechanical model.

40. The non-transitory computer-readable storage medium of claim 35, wherein: The treatment target includes a three-dimensional 3D model.

41. A computer-implemented method comprising: receiving orthognathic patient data, wherein the patient data includes information representing an initial state of the patient; receiving a plurality of treatment parameters, wherein the plurality of treatment parameters are based on treatment goals for the patient; generating an output state for the patient based on the initial state of the patient and the plurality of treatment parameters; calculating an objective function based on the patient's output state and the patient's treatment goal; and One or more of the treatment parameters are updated based on the objective function.

42. The computer-implemented method of claim 41 , wherein: The orthognathic patient data includes one or more surface images of the patient's face.

43. The computer-implemented method of claim 41 , wherein: The orthognathic patient data includes one or more three-dimensional models of the patient's face.

44. The computer-implemented method of claim 41 , wherein: The orthognathic patient data includes one or more three-dimensional models of the patient's jaws.

45. The computer-implemented method of claim 41 , wherein: The treatment parameters include one or more facial landmarks associated with the patient.

46. ​​The computer-implemented method of claim 41 , wherein: Updating the one or more treatment parameters further comprises: The one or more treatment parameters are identified based on a derivative of the objective function.

47. The computer-implemented method of claim 41, wherein: Generating the output status of the patient further comprises: The orthognathic patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a numerical model.

48. The computer-implemented method of claim 47, wherein: The numerical model includes a biomechanical model.

49. The computer-implemented method of claim 41, wherein: The output state of the patient comprises a three-dimensional 3D mesh.

50. A system comprising: Memory; as well as A processor is operably coupled to the memory, the processor being configured to: receiving orthognathic patient data, the orthognathic patient data including information representing an initial state of the patient; receiving a plurality of treatment parameters based on a treatment goal for the patient; generating an output state for the patient based on the initial state of the patient and the plurality of treatment parameters; calculating an objective function based on the patient's output state and the patient's treatment goal; as well as One or more of the treatment parameters are updated based on the objective function.

51. The system of claim 50, wherein: The orthognathic patient data includes one or more surface images of the patient's face.

52. The system of claim 50, wherein: The orthognathic patient data includes one or more three-dimensional models of the patient's face.

53. The system of claim 50, wherein: The orthognathic patient data includes one or more three-dimensional models of the patient's jaws.

54. The system of claim 50, wherein: The treatment parameters include one or more facial landmarks associated with the patient.

55. The system of claim 50, wherein: The processor is further configured to: The one or more treatment parameters are identified based on a derivative of the objective function.

56. The system of claim 50, wherein: The processor is further configured to: The orthognathic patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model comprises a numerical model.

57. The system of claim 56, wherein: The numerical model includes a biomechanical model.

58. The system of claim 50, wherein: The output state of the patient comprises a three-dimensional 3D mesh.

59. A computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform operations comprising: receiving orthognathic patient data, the orthognathic patient data including information representing an initial state of the patient; receiving a plurality of treatment parameters based on a treatment goal for the patient; generating an output state for the patient based on the initial state of the patient and the plurality of treatment parameters; calculating an objective function based on the patient's output state and the patient's treatment goal; as well as One or more of the treatment parameters are updated based on the objective function.

60. The computer readable medium of claim 59, wherein: Generating the output status of the patient includes: The orthognathic patient data and the plurality of treatment parameters are provided to a simulation model, wherein the simulation model includes a numerical model that generates output states for the patient.

Images