Exploration of Pareto optimal radiotherapy plan

Abstract

Systems and methods for exploring and adjusting radiation therapy plans are disclosed. Exemplary operations for radiotherapy planning include obtaining a plurality of solutions to a radiotherapy problem (e.g., a Pareto optimal solution), exploring the plurality of solutions to identify an additional solution in a substantive manifold space (e.g., an exploration of a Pareto surface), and generating therapy planning parameters based on the additional solution for use in radiotherapy. In one example, exploring the plurality of solutions includes: establishing a sub-manifold space from a manifold space, the sub-manifold space representing the plurality of solutions in a dimension less than a weight; generating an additional weight set in the sub-manifold space based on a derivative of a first-order optimality condition of the radiotherapy problem, the derivative being determined with respect to the weight; and navigating in the sub-manifold space to obtain an additional solution, the additional solution corresponding to one of the additional sets of weights.

Description

Technical Field

[0001] Embodiments of the present disclosure generally relate to processing and optimization techniques used in conjunction with radiation therapy planning and treatment systems. Specifically, the present disclosure relates to methods and uses of specific computing hardware configurations to navigate multiple Pareto-optimal radiation therapy plans. Background Art

[0002] Radiation therapy (or "radiotherapy") can be used to treat cancer or other diseases in mammalian (e.g., human and animal) tissue. One such radiotherapy technique is provided using a gamma knife, whereby a patient is irradiated with a large number of low-intensity gamma rays that are focused at a target (e.g., a tumor) with high intensity and precision. Another such radiotherapy technique is provided using a linear accelerator (Linac), whereby a tumor is irradiated with high-energy particles (e.g., electrons, protons, ions, high-energy photons, etc.). The placement and dose of the radiation beam must be precisely controlled to ensure that the tumor receives the prescribed radiation and that the radiation beam is placed to minimize damage to surrounding healthy tissue, often referred to as organs at risk (OARs).

[0003] In radiotherapy, treatment plans are typically developed by solving an optimization problem that balances various conflicting objectives, including, for example, high doses to the target tumor, protection of healthy tissue, and other treatment complexities. The optimization problem is therefore a multi-criteria optimization (MCO) problem. The number of competing objectives typically ranges from a few to about 10, depending on the tumor location. Typically, different criteria are combined using a weighted sum, where each weight determines the relative importance of that criterion. For convex optimization problems, the optimization attempts to determine an exact approximation of a set of Pareto optimal radiotherapy plans. In this context, a Pareto optimal plan refers to a radiotherapy plan in which an improvement in any one criterion does not lead to a deterioration in another criterion.

[0004] Finding acceptable weights to develop a Pareto-optimal radiation therapy plan is often a manual and tedious trial-and-error process, especially because evaluating a single choice of parameters requires solving the full optimization problem for the radiation therapy. Solving the full optimization problem for a single plan can take anywhere from a few seconds to an hour (depending on its application) to calculate and evaluate parameter combinations. Furthermore, even if a set of Pareto-optimal plans are developed for a particular radiation therapy, it can take a significant amount of time and effort to identify, select, and refine the best plan for implementation. Summary of the invention

[0005] Various embodiments, methods, systems, and computer-readable media are provided for evaluating and optimizing radiation therapy plans, including exploring or directing the results of solving a radiation therapy problem (e.g., a radiation therapy problem that can be adjusted via specific parameterized criteria and solved using an optimization method). The techniques can be used to evaluate the results of an optimization method that has produced multiple possible solutions. Such evaluations can then be used to identify specific parameter selections or combinations of parameters that result in a satisfactory radiation therapy plan.

[0006] As described herein, the radiation therapy problem can be expressed as a multi-criteria optimization (MCO) problem, where the parameters correspond to clinical preferences. The multiple solutions generated for this radiation therapy problem can constitute multiple Pareto optimal plans. The Pareto optimal plan used in this article refers to an optimized plan that improves any one criterion without deteriorating another criterion. The set of all Pareto optimal plans constitutes a Pareto surface (also called a Pareto boundary).

[0007] An extended method for radiation therapy plan optimization is provided below, which uses a special technique to navigate the Pareto surface of the solution to the MCO problem. In one example, the Pareto surface can be explored starting from a Pareto optimal initial point using a submanifold with a lower dimension of the weight space (the Pareto surface is path-connected). The solution to the MCO problem can then be estimated based on the derivative of the first-order optimality condition.

[0008] In various examples, the operations for the radiation therapy planning may include a process for obtaining a plurality of solutions to a radiation therapy problem defined in a manifold space for providing radiation therapy, exploring the plurality of solutions to identify additional solutions in a submanifold space, and generating treatment plan parameters based on the additional solutions, wherein the treatment plan parameters are provided for delivering the radiation therapy via a radiation therapy machine in the treatment plan. The radiation therapy problem is a multi-criteria optimization problem, and each of the plurality of solutions has a plurality of weights for adjusting a plurality of criteria of the multi-criteria optimization problem.

[0009] In one example, additional solutions are identified and explored based on: (a) establishing a submanifold space from the manifold space, the submanifold space representing multiple solutions in fewer dimensions than the multiple weights; generating additional sets of weights in the submanifold space based on derivatives of a first-order optimality condition for the radiation therapy problem, such derivatives being determined relative to the multiple weights; and (c) performing navigation in the submanifold space to obtain additional solutions corresponding to one of the additional weight sets. For example, navigation in the submanifold space can be performed using a one-dimensional path. Likewise, in one example, navigation of the one-dimensional path can be performed starting from a Pareto optimal initial point because navigation is simplified to finding a one-dimensional path by solving an ordinary differential equation based on directional derivatives.

[0010] In another example, the identification of additional solutions solves a boundary value problem, providing an initial point and a final point for navigation of a one-dimensional path, as the navigation interpolates between the initial point and the final point. Also in another example, the identification of additional solutions solves an initial value problem, providing an initial point for navigation of a one-dimensional path, as the navigation operates until a predetermined stopping condition is satisfied. The predetermined stopping condition can be at least one of: a maximum acceptable degradation of a time, distance, or clinical indicator (e.g., using a simultaneous localization and mapping (SLAM) method to improve a first clinical indicator while minimizing degradation of a second clinical indicator, thereby determining degradation).

[0011] In other examples, the additional set of weights in the submanifold space corresponds to a level set of optimality conditions for the radiation therapy problem. Also in other examples, the derivative is provided by automatic differentiation. Also in other examples, at least a portion of the manifold space has a non-differentiable portion. In this non-differentiable case, the non-differentiable portion of the manifold space can be smoothly approximated using a least squares method; or, the constraint problem can be converted to an unconstrained problem using a barrier formulation by including the constraint as a term in the objective function and associating a penalty for violating the constraint.

[0012] In other examples, the multiple criteria of the radiation therapy problem correspond to clinical preferences. For example, the parameters of the multiple criteria may relate to specific anatomical regions for receiving radiation therapy from a radiation therapy machine. Likewise, the multiple criteria may include the definition of one or more organs at risk and one or more target regions.

[0013] In other examples, a representation of potential solutions, additional solutions, or other aspects of solution exploration for a radiation therapy problem may be generated and output. The graphical representation may include a display provided from a graphical user interface having functionality for configuring a treatment plan, and an information display in the graphical user interface that provides information associated with the additional solution. The graphical user interface or other user interfaces may also receive a user or automatically selected additional solution. In a further example, the selected additional solution provides a hot start for identifying a final solution for the treatment plan, and one or more additional optimizations of the selected additional solution may be received before generating treatment plan data from the final solution.

[0014] The treatment plan parameters for radiotherapy include a set of treatment delivery parameters corresponding to the capabilities of the radiotherapy machine. In one example, the treatment plan is used to provide radiotherapy with a gamma knife, and the set of treatment delivery parameters includes a set of isocenters for delivering the radiotherapy. For example, the set of treatment delivery parameters also includes the timing for delivering the radiotherapy, and the collimator sequence for delivering the radiotherapy with the gamma knife. As another example, the treatment plan is used to provide volumetric arc modulated therapy (VMAT) or intensity modulated radiation therapy (IMRT) to the radiotherapy (e.g., using a linear accelerator radiotherapy machine), and the set of treatment delivery parameters includes one or more of the following: a set of arc control points for one or more arcs, a fluence field, a gantry speed, and a dose rate along the one or more arcs.

[0015] In another example, following the operations described above, delivery of radiation therapy may be induced or affected using a plurality of radiation therapy beams from a radiation therapy machine based on treatment planning parameters or other data generated or identified as discussed herein.

[0016] The above invention content is intended to provide an overview of the subject matter of this patent application. This disclosure is not intended to provide an exclusive or exhaustive explanation of the subject matter of the invention. This disclosure contains specific embodiments to provide further information about this patent application. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In the drawings, which are not necessarily drawn to scale, like numerals describe substantially similar components in the several views. Like numerals with different letter suffixes represent different instances of substantially similar components. The drawings generally illustrate various embodiments discussed in this document by way of example and not limitation.

[0018] Figure 1 A radiation therapy system according to some examples is shown.

[0019] Figure 2A A radiation therapy system according to some examples is shown having an output configured to provide a therapy beam.

[0020] Figure 2B A system according to some examples is shown that includes a combined radiation therapy system and an imaging system, such as a cone beam computed tomography (CBCT) imaging system.

[0021] Figure 3 A partial cross-sectional view of a system including a combined radiation therapy system and an imaging system, such as a magnetic resonance (MR) imaging (MRI) system, is shown according to some examples.

[0022] Figure 4An example of a Leksell Gamma Knife radiotherapy device is shown, according to some examples.

[0023] Figure 5 A radiation therapy planning workflow according to some examples is shown.

[0024] Figure 6 A further workflow for identifying and exploring additional Pareto optimal solutions according to some examples is shown.

[0025] Figure 7 A representation of exploration in a Pareto surface according to some examples is shown.

[0026] Figure 8A-8D Representations of Pareto surfaces and submanifolds according to some examples are shown.

[0027] Fig. 9 A flow chart illustrating a method of radiation therapy planning according to some examples is shown.

[0028] Fig.10 An exemplary block diagram of a machine is shown upon which one or more of the methods discussed herein may be implemented. DETAILED DESCRIPTION

[0029] In the following detailed description, reference is made to the accompanying drawings that form a part of the detailed description, and the detailed description is shown by way of a diagrammatic specific embodiment in which the present disclosure can be practiced. These embodiments, also referred to herein as "examples", are described in sufficient detail to enable those skilled in the art to practice the present disclosure, and it should be understood that the embodiments may be combined or other embodiments may be utilized and structural, logical, and electrical changes may be made without departing from the scope of the present disclosure. Therefore, the following detailed description is not restrictive, and the scope of the present disclosure is limited by the attached aspects and their equivalents.

[0030] Various implementations of planning and optimization techniques that can be used for radiotherapy or radiosurgery applications are discussed below. Such techniques can be used to find and evaluate new solutions to radiotherapy problems that may correspond to different parameter selections and values ​​(weights) for such parameterized criteria. Based on the exploration, selection, and refinement of these parameters and parameter values, treatment planning data and a unique treatment plan implemented with a radiotherapy machine can be generated.

[0031] In the example discussed herein, multiple solutions to a radiation therapy optimization problem can be evaluated during the process of "surfing" the Pareto surface of the solution. Starting from the Pareto optimal initial point, the following technique provides an exploration of the Pareto surface using a low-dimensional submanifold in the weight space, which is path-connected. Other solutions to the multi-criteria radiation therapy optimization problem represented by the submanifold can be estimated based on the derivative of the first-order optimality condition. The first-order optimality condition is a necessary condition for a point to become an optimal point. For unconstrained optimization problems, stationarity (i.e., the gradient of the objective function is zero) is a first-order optimality condition. For constrained optimization problems, meeting the Karush-Kuhn-Tucker (KKT) condition is a first-order optimality condition.

[0032] The currently disclosed exploration and generation of solutions to radiation therapy problems offer significant technical and clinical advantages. Technical advantages include reduced computational processing time for generating new radiation therapy plans, enhanced computational solutions to radiation therapy planning optimization problems, and concomitant improvements in processing, memory, and network resources used to generate or perform radiation therapy. Related clinical advantages include potentially improved plan quality and reduced treatment time.

[0033] Capturing the preferences and requirements of a treatment planner in a single mathematical expression is extremely challenging, if not impossible. Even if such a model could be obtained, it would almost certainly be difficult to optimize. Instead, the following uses a set of well-behaved (e.g., convex) surrogate functions as building blocks that, when appropriately combined, result in treatment plans that meet the treatment planner's actual preferences and requirements. To ensure that such plans exist, the surrogate functions are typically chosen to be associated with specific clinical considerations and to collectively span the relevant range of treatment plans. However, due to the mismatch between actual clinical considerations and the surrogate functions, and because true clinical preferences are often unknown, it is also unknown how to weight the surrogate functions to find a clinically satisfactory plan.

[0034] Therefore, it is necessary to evaluate different weighting settings. This is often a tedious and time-consuming process, as optimization must be performed at each step. As a result, multiple attempts are unlikely to be made with the risk that the plan ultimately chosen may not be as good as what could have been achieved. In fact, multiple studies have found that treatment plan quality is highly dependent on the time spent and experience level of the individual treatment planner.

[0035] Generating a large collection of plans (i.e., an approximation of the Pareto surface) during a single optimization, and providing tools to navigate on this potentially high-dimensional surface, increases the likelihood of finding a better plan for the patient in less time than would be possible with an optimization process involving manual intervention. In addition, the approximation of the Pareto surface provides the treatment planner with an overview of the achievable tradeoffs, which makes it easier to identify promising areas for further exploration, as well as to know when to terminate the search for a better plan. Both aspects can reduce planning time and potentially improve average plan quality.

[0036] The techniques used to select a clinical plan from a Pareto approximation are also important. Discretely selecting a plan from precomputed plans requires a very dense representation that is not feasible in practice. A coarser representation is sufficient if continuous interpolation between discrete solution sets is allowed. This form of continuous interpolation is often referred to as navigation. Plans optimized based on a weighted sum of objective (surrogate) functions can be shown to be Pareto optimal plans for those objectives, but not necessarily with respect to actual clinical considerations. The navigation tools provided by the following techniques allow clinicians to explore and prioritize plans based on actual clinical considerations, thereby determining the best plan more quickly and reliably.

[0037] As described below, the following techniques can be used to represent and explore a set of treatment plans, such as a Pareto surface of Pareto-optimal treatment plans, allowing the discovery of new treatment parameter weights that can be identified and selected for use. Existing methods for representing and exploring a set of representative treatment plans using conventional processing techniques often result in significant computational time and latency. For example, some existing methods involve pre-computing a set of representative Pareto-optimal treatment plans and selecting the most appropriate plan from them. The number of representative plans required to traverse the Pareto surface for a Gamma Knife treatment is typically on the order of hundreds or thousands of plans, which means that very long computational times (typically overnight runs) are required to generate a complete representation of possible solutions. Some existing methods also attempt to select and combine weights from different solutions, such as using interpolation schemes. However, the evaluation and exploration of such solutions is quite computationally intensive and requires a large number of samples. In contrast, techniques for efficient Pareto navigation are discussed below, as well as the interactive steps taken to identify a satisfactory compromise given a starting set of Pareto-optimal plans.

[0038] The following paragraphs provide an overview of example radiation therapy system implementations and treatment use cases (see Figure 2A , Figure 2B , Figure 3 and Figure 4 ), including the use of computing systems and hardware implementations (reference Figure 1 and 10 ). We then move on to discuss an example treatment plan exploration and development workflow (ref. Figure 5 and Figure 6 ), which includes ways to explore representations of Pareto surfaces (ref. Figure 7 and a representation of the Pareto surface in Figure 8). Finally, a discussion of radiation therapy planning is provided (ref. Fig. 9 ), which shows an end-to-end approach to generating an optimized treatment plan.

[0039] Figure 1 An exemplary radiation therapy system 100 suitable for performing radiation therapy planning processing operations using one or more of the methods discussed herein is shown. These radiation therapy planning processing operations are performed to enable the radiation therapy system 100 to provide radiation therapy to a patient based on captured medical imaging data and treatment dose calculations or specific aspects of radiation therapy machine configuration parameters. Specifically, the following processing operations may be implemented as part of the radiation therapy planning logic 120 for exploring and developing parameter weights for radiation therapy plans. However, it will be appreciated that many variations and use cases of the following planning logic 120 and optimization operations may be provided, such as in response to data validation, visualization, and other medical evaluation and diagnostic operations.

[0040] The radiation therapy system 100 includes a radiation therapy treatment computing system 110 that hosts the radiation therapy planning logic 120. The radiation therapy treatment computing system 110 can be connected to a network (not shown), and the network can be connected to the Internet. For example, the network can connect the radiation therapy treatment computing system 110 with one or more private and / or public medical information sources (e.g., a radiology information system (RIS), a medical record system (e.g., an electronic medical record (EMR) / electronic health record (EHR) system), an oncology information system (OIS)), one or more image data sources 150, an image acquisition device 170 (e.g., an imaging modality), a treatment device 180 (e.g., a radiation therapy device), and a treatment data source 160.

[0041] As an example, the radiation therapy treatment computing system 110 may be configured to receive a treatment objective for a subject (e.g., one or more MR images) and generate a radiation therapy plan by executing instructions or data from the radiation therapy planning logic 120 as part of generating an operation to be used by the treatment device 180 and / or for output on the device 146. In one embodiment, the radiation therapy planning logic 120 generates a solution to an optimization problem and then uses one or more solutions to generate a radiation therapy plan. The radiation therapy planning logic 120 solves the radiation therapy optimization problem by estimating the optimization variables of the received optimization problem. The optimization problem is then solved using an optimization problem solver, generating a plurality of possible solutions (e.g., Pareto optimal solutions). The possible solutions to the radiation therapy optimization problem are then further explored and refined using the techniques discussed herein.

[0042] Exemplary optimization problem solvers include, for example, simplex method, interior point method, Newton method, quasi-Newton method, Gauss-Newton method, Levenberg-Marquardt method, linear least squares method, gradient descent method, projected gradient method, conjugate gradient method, augmented Lagrangian method, Nelder-Mead method, branch and bound method, cutting plane method, simulated annealing and / or sequential quadratic programming. In other examples, the optimization problem solver may include using alternating direction multiplier method (ADMM) applied to parallel processing circuits. Another type of optimization solver uses a machine learning model to predict the solution or some intermediate results of the calculation, and can be combined or interleaved with steps performed according to optimization methods not based on machine learning.

[0043] The general radiation therapy planning optimization problem can be defined as Equation 1:

[0044]

[0045] subject to x∈Ω

[0046] (Equation 1)

[0047] in is the objective function, x∈X is the decision variable (also called optimization variable), is the set of feasible variables. In general, the function f can be nonlinear and the set Ω is nonconvex. Optimization problems are usually solved using some form of iterative scheme. For example, if f is smooth and convex, and Ω is convex, then a projected gradient scheme can be used to solve equation (1), which reads as follows:

[0048]

[0049] where projΩ:X→X is the projection onto Ω, is the step size, and X→X is the gradient. Such algorithms are often provably convergent (i.e., given enough time (and the right choice of parameters), the algorithm will converge to a minimum). Some of these algorithms may require hundreds (or even thousands) of iterations to achieve approximate convergence. Since each step can be computationally expensive, this can mean a run time of minutes or even hours. However, using the techniques discussed in this article, it is possible to explore such solutions and create additional parameter values ​​based on them.

[0050] The radiation therapy treatment computing system 110 may include a processing circuit 112, a memory 114, a storage device 116, and other hardware and software operable features, such as a user interface 142, a communication interface (not shown), etc. The storage device 116 may store temporary or non-temporary computer executable instructions, such as an operating system, radiation therapy plans, training data, software programs (e.g., image processing software, image or anatomical visualization software, artificial intelligence (AI) or machine learning (ML) implementations and algorithms (e.g., provided by deep learning models, ML models, neural networks (NNs), etc.)), and any other computer executable instructions executed by the processing circuit 112.

[0051] In one example, the processing circuit 112 may include a processing device, such as one or more general-purpose processing devices (e.g., a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), an accelerated processing unit (APU), etc.). More specifically, the processing circuit 112 may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor that implements other instruction sets, or a processor that implements a combination of instruction sets. The processing circuit 112 may also be implemented by one or more special-purpose processing devices (e.g., an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a system on a chip (SoC), etc.).

[0052] As will be appreciated by those skilled in the art, in some examples, processing circuit 112 may be a dedicated processor rather than a general purpose processor. Processing circuit 112 may include one or more known processing devices, such as Intel TM Manufacturing Pentium TM , Core TM , Xeon TM or family of microprocessors; AMD TM Made by Turion TM , Athlon TM 、Sempron TM 、Opteron TM FX TM 、Phenom TM family of processors; or any processor manufactured by Sun Microsystems. Processing circuitry 112 may also include a graphics processing unit, such as an Nvidia TM Manufactured series of GPUs; Intel TM GMA, Iris TM series GPU; or AMD TM Made by RadeonTM The processing circuit 112 may also include an accelerated processing unit, such as Intel TM Xeon Phi manufactured TM The disclosed embodiments are not limited to any type of processor that is otherwise configured to meet the computational requirements of identifying, analyzing, maintaining, generating and / or providing large amounts of data or manipulating such data to perform the methods disclosed herein. Furthermore, a "processor" may include more than one physical (circuit-based) or software-based processor (e.g., a processor of a multi-core design or even multiple processors each having a multi-core design).

[0053] Other embodiments of processing circuitry 112 may include processors or processing units arranged in a parallel processing configuration. For example, a group of graphics processing units (e.g., Nvidia TM Manufactured series of GPUs; Intel TM GMA, Iris TM series GPU; or AMD TM Made by Radeon TM A GPU (a series of GPUs) can be arranged to perform highly parallel or repetitive computing tasks simultaneously. Other specialized parallel processing units or hardware capable of performing multiple calculations simultaneously may also be deployed. A GPU may include a single GPU "device" or "system" that operates or coordinates numerous (e.g., tens, hundreds, or thousands) subprocessors; use of a GPU may include use of a single GPU device or system that uses each of its numerous subprocessors to process a corresponding set of parameters.

[0054] Processing circuit 112 may execute sequences of transient or non-transitory computer program instructions stored in memory 114 and accessed from storage device 116 to perform various operations, processes, and methods that will be explained in more detail below. It should be understood that any component in system 100 may be implemented separately and operated as a standalone device, and may be coupled to any other component in system 100 to perform the techniques described in the present disclosure.

[0055] The memory 114 may include read-only memory (ROM), phase-change random access memory (PRAM), static random access memory (SRAM), flash memory, random access memory (RAM), dynamic random access memory (DRAM) (e.g., synchronous dynamic random access memory (SDRAM)), electrically erasable programmable read-only memory (EEPROM), static memory (e.g., flash memory, flash disk, static random access memory), and other types of random access memory, cache, registers, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, cassettes, other magnetic storage devices, or any other non-transitory medium that can be used to store information (including images, training data, one or more ML models or technique parameters, data, or transient or non-transitory computer executable instructions (e.g., stored in any format)) that can be accessed by the processing circuit 112 or any other type of computer device. For example, computer program instructions can be accessed by the processing circuit 112, read from ROM or any other suitable memory location, and loaded into RAM for execution by the processing circuit 112.

[0056] The storage device 116 may constitute a driver unit including a transitory or non-transitory machine-readable medium having stored thereon one or more sets of transitory or non-transitory instructions and data structures (e.g., software) that embody or utilize any one or more of the methods or functionality described herein (including, in various examples, the radiation therapy planning logic 120 and the user interface 142). During execution by the radiation therapy treatment computing system 110, the instructions may also reside, in whole or in part, within the memory 114 and / or within the processing circuitry 112, wherein the memory 114 and the processing circuitry 112 also constitute transitory or non-transitory machine-readable media.

[0057] The memory 114 and the storage device 116 may constitute a non-transitory computer-readable medium. For example, the memory 114 and the storage device 116 may store or load temporary or non-transitory instructions for one or more software applications on a computer-readable medium. The software applications stored or loaded by the memory 114 and the storage device 116 may include, for example, operating systems for ordinary computer systems and for software-controlled devices. The radiation therapy treatment computing system 110 may also operate various software programs including software code for implementing the radiation therapy planning logic 120 and the user interface 142. In addition, the memory 114 and the storage device 116 may store or load the entire software application, a portion of the software application, or code or data associated with the software application, which may be executed by the processing circuit 112. In another example, the memory 114 and the storage device 116 may store, load, and manipulate one or more radiation therapy plans, imaging data, segmentation data, treatment visualization, histograms or measurements, one or more artificial intelligence model data (e.g., weights and parameters of one or more artificial intelligence models), training data, label and mapping data, etc. It is contemplated that the software program may be stored not only on storage device 116 and memory 114, but also on removable computer media such as a hard drive, computer disk, CD-ROM, DVD, Blu-ray DVD, USB flash drive, SD card, memory stick, or any other suitable medium. Such software programs may also be communicated or received over a network.

[0058] Although not depicted, the radiation therapy treatment computing system 110 may include communication interfaces, network interface cards, and communication circuits. For example, example communication interfaces may include network adapters, cable connectors, serial connectors, USB connectors, parallel connectors, high-speed data transmission adapters (e.g., such as fiber optics, USB 3.0, Thunderbolt, etc.), wireless network adapters (e.g., such as IEEE 802.11 / Wi-Fi adapters), telecommunications adapters (e.g., communicating with 3G, 4G / LTE, 5G, etc. networks), etc. Such communication interfaces may include one or more digital and / or analog communication devices that allow the machine to communicate with other machines and devices (e.g., remotely located components) via a network. The network may provide the functionality of a local area network (LAN), a wireless network, a cloud computing environment (e.g., software as a service, platform as a service, infrastructure as a service, etc.), a client-server, a wide area network (WAN), etc. For example, the network may be a LAN or a WAN, which may include other systems (including additional image processing computing systems or image-based components associated with medical imaging or radiation therapy operations).

[0059] In one example, the radiation therapy treatment computing system 110 may obtain image data 152 from an image data source 150 (e.g., an MR image) for hosting on the storage device 116 and the memory 114. In yet another example, a software program may substitute a function of a patient image, such as a signed distance function or a processed version of the image that emphasizes certain aspects of the image information. The radiation therapy treatment computing system 110 may obtain image data 152 from or transmit it to the image data source 150. In other examples, the treatment data source 160 receives or updates planning data 162 as a result of a treatment plan generated by the radiation therapy planning logic 120. The image data source 150 may also provide or host imaging data for the radiation therapy planning logic 120.

[0060] In one example, the computing system 110 can communicate with one or more treatment data sources 160, input devices 148, and other data sources to generate optimization parameters and weights for multiple radiation therapy planning optimization problems. These optimization variables and parameters are then evaluated and refined to determine a specific solution to the radiation therapy problem. It should be understood that the results of solution generation and solution exploration may approximate the final implemented solution, but may provide a warm start for identifying the true solution. Therefore, multiple iterations of generation, identification, exploration, and refinement may be involved before a solution is used in radiation therapy.

[0061] The processing circuitry 112 may be communicatively coupled to the memory 114 and the storage device 116, and the processing circuitry 112 may be configured to execute computer executable instructions stored on the memory 114 or the storage device 116. In particular, the radiation therapy planning logic 120 receives an optimization problem derived from radiation therapy parameters. The processing circuitry 112 may utilize a software program or implementation to explore and develop parameters for delivering radiation therapy doses to a patient as part of developing an optimal solution to the radiation therapy problem as discussed herein. In addition, such software program or implementation may utilize the radiation therapy planning logic 120 to generate new or updated treatment plan parameters using the techniques further discussed herein for deployment to the treatment data source 160 and / or presentation on the output device 146. Subsequently, the processing circuitry 112 may transmit the new or updated treatment plan details to the treatment device 180 via the communication interface and network, wherein the radiation therapy plan will be used to treat the patient with radiation via the treatment device 180, consistent with the results of the radiation therapy planning logic 120 (e.g., according to the process discussed below).

[0062] In one example, the image data 152 for defining a radiation therapy question or indicating an anatomical region of a patient may include one or more of the following: an MRI image (e.g., two-dimensional (2D) MRI, three-dimensional (3D) MRI, two-dimensional flow MRI, four-dimensional (4D) MRI, four-dimensional volume MRI, four-dimensional cine MRI, etc.), a functional MRI image (e.g., fMRI, DCE-MRI, diffusion MRI), a computed tomography (CT) image (e.g., 2D-CT, 2D cone beam CT, 3D-CT, 3D cone beam CT, 4D-CT, 4D cone beam CT), an ultrasound image (e.g., 2D ultrasound, 3D ultrasound, 4D ultrasound), a positron emission tomography image (PET), an X-ray image, a fluoroscopic image, a radiation therapy portal image, a single photon emission computed tomography image (SPECT), a computer-generated synthetic image (e.g., a pseudo CT image), etc. In addition, the image data 152 may also include or be associated with medical image processing data (e.g., training images, ground truth images, contour images, and dose images). In other examples, equivalent representations of anatomical regions may be represented in non-image formats (eg, coordinates, maps, etc.).

[0063] In one example, the image data 152 may be received from the image acquisition device 170 and stored in one or more of the following: an image data source 150 (e.g., a picture archiving and communication system (PACS), a vendor neutral archive (VNA), a medical record or information system, a data warehouse, etc.), etc. Thus, the image acquisition device 170 may include an MRI imaging device, a CT imaging device, a PET imaging device, an ultrasound imaging device, a fluoroscopy device, a SPECT imaging device, an integrated linear accelerator and MRI imaging device, a CBCT imaging device, or other medical imaging device for acquiring medical images of a patient. The image data 152 may be received and stored in any type of data or any type of format (e.g., in a digital imaging and communications in medicine (DICOM) format) that the image acquisition device 170 and the radiation therapy treatment computing system 110 can use to perform operations consistent with the disclosed embodiments. In addition, in some examples, the models discussed herein may be trained to process the original image data format or its derivatives.

[0064] In one example, the image acquisition device 170 can be integrated with the treatment device 180 into a single device (e.g., an MRI device combined with a linear accelerator, also referred to as an "MRI-Linac"). For example, such an MRI-Linac can be used to determine the location of a target organ or target tumor in a patient's body so that the radiotherapy can be accurately guided to a predetermined target point according to the radiotherapy plan. For example, the radiotherapy plan can provide information about the specific radiation dose to be applied to each patient. The radiotherapy plan may also include other radiotherapy information, including control points of the radiotherapy device, such as bed position, beam intensity, beam angle, dose histogram volume information, the number of radiation beams used during treatment, the dose of each beam, etc.

[0065] The radiation therapy treatment computing system 110 can communicate with an external database via a network to send / receive various types of data related to image processing and radiation therapy operations. For example, the external database may include machine data (including equipment constraints) that provides information related to the treatment device 180, the image acquisition device 170, or other machines related to the radiation therapy or medical procedure. The machine data information (e.g., control points) may include radiation beam size, arc placement position, beam on and off duration, machine parameters, segments, multi-leaf collimator (MLC) configuration, gantry speed, MRI pulse sequence, etc. The external database may be a storage device and may be equipped with an appropriate database management software program. In addition, such a database or data source may include multiple devices or systems located in a centralized or distributed manner.

[0066] The radiation therapy treatment computing system 110 may collect and acquire data and communicate with other systems via a network using one or more communication interfaces that are communicatively coupled to the processing circuitry 112 and the memory 114. For example, the communication interface may provide a communication connection between the radiation therapy treatment computing system 110 and radiation therapy system components (e.g., allowing data to be exchanged with external devices). For example, in some examples, the communication interface may have appropriate interface circuitry from the output device 146 or the input device 148 to connect to the user interface 142, which may be a hardware keyboard, keypad, or touch screen through which a user may input information into the radiation therapy system.

[0067] As an example, the output device 146 may include a display device that outputs a representation of the user interface 142 and one or more aspects, visualizations, or representations of a medical image, a treatment plan, a solution, or a solution space, and the status of the training, generation, validation, or implementation of these plans. The output device 146 may include one or more display screens that display medical images, interface information, treatment plan parameters (e.g., contours, doses, beam angles, labels, maps, etc.), treatment plans, targets, localization targets, and / or tracking targets, or any related information to a user. In a specific example, the user interface 142 includes a solution navigation interface that allows operation and interaction with various aspects of the solution exploration 136 (e.g., using a graphical interface discussed below).

[0068] The input device 148 connected to the user interface 142 can be a keyboard, keypad, touch screen, or any type of device that a user can use with the radiation therapy system 100. Alternatively, the features of the output device 146, input device 148, and user interface 142 can be integrated into a single device, such as a smartphone or tablet (e.g., Apple Lenovo Samsung wait).

[0069] In addition, any and all components of the radiation therapy system can be implemented as virtual machines (e.g., via a virtualization platform such as VMWare, Hyper-V, etc.) or stand-alone devices. For example, a virtual machine can be software that functions as hardware. Thus, a virtual machine can include at least one or more virtual processors, one or more virtual memories, and one or more virtual communication interfaces that together function as hardware. For example, the radiation therapy processing computing system 110, the image data source 150, or similar components can be implemented as virtual machines or within a cloud-based virtualization environment.

[0070] The image acquisition device 170 can be configured to acquire one or more images of the patient's anatomical structure of the region of interest (e.g., target organ, target tumor, or both). Each image (usually a 2D image or slice) can include one or more parameters (e.g., 2D slice thickness, orientation, and position, etc.). In one example, the image acquisition device 170 can acquire 2D slices in any orientation. For example, the orientation of the 2D slice can include sagittal orientation, coronal orientation, or axial orientation. The processing circuit 112 can adjust one or more parameters, such as the thickness and / or orientation of the two-dimensional slice to include the target organ and / or target tumor. In one example, the 2D slice can be determined based on information (e.g., 3D-CBCT or CT or MRI volume). When the patient is receiving radiation therapy (e.g., when using the treatment device 180), the image acquisition device 170 can acquire these two-dimensional slices in "near real time" ("near real time" means collecting data in at least milliseconds or less).

[0071] The radiation therapy planning logic 120 in the radiation therapy treatment computing system 110 implements a radiation therapy optimization workflow 130 and a treatment planning workflow 140. The radiation therapy optimization workflow 130 may implement optimization operations for identifying and developing parameters of a radiation therapy plan, while the treatment planning workflow 140 may implement operations for implementing these parameters of a radiation therapy plan for a specific purpose through a radiation therapy machine. In a specific example, the radiation therapy optimization workflow 130 performs radiation therapy problem processing 132 to obtain and identify an optimization problem, performs Pareto optimal solution generation 134 to identify and output Pareto optimal solutions to a multi-criteria optimization problem, and performs Pareto optimal solution exploration 136 to explore these solutions in a reduced submanifold space and generate additional solutions. Reference is made below to Figure 6 and Figure 7 Further details of the radiation therapy optimization workflow 130 are provided, including the use of various methods to explore and generate unique solutions to radiation therapy problems. Figure 5 and Fig. 9 Further details of the treatment planning workflow 140 are provided, showing how unique problem solutions are generated and implemented into the treatment planning process.

[0072] Figure 2A A radiotherapy device 202 is shown, which may include a radiation source, such as an x-ray source or a linear accelerator, a couch 216, an imaging detector 214, and a radiotherapy output 204. The radiotherapy device 202 may be configured to emit a radiation beam 208 to provide treatment to a patient. The radiotherapy output 204 may include one or more attenuators or collimators, such as an MLC. The MLC may be used to shape, direct, or modulate the intensity of a radiotherapy beam to a specific target in a patient's body. For example, the leaves of the MLC may be automatically positioned to define an aperture that is close to a cross-section or projection of a tumor and cause modulation of the radiotherapy beam. For example, the leaves may include a metal plate, such as tungsten, with the long axis of the plate oriented parallel to the beam direction and having an end oriented orthogonal to the beam direction. In addition, the "state" of the MLC may be adaptively adjusted during the radiotherapy process, such as to establish a treatment beam that is closer to the shape or position of a tumor or other target.

[0073] Re-reference Figure 2A, according to the radiation treatment plan, the patient may be positioned in the area 212 and supported by the treatment couch 216 to receive the radiation treatment dose. The radiation treatment output 204 may be mounted or attached to a gantry 206 or other mechanical support. When the couch 216 is inserted into the treatment area, one or more chassis motors (not shown) may rotate the gantry 206 and the radiation treatment output 204 about the couch 216. In one example, the gantry 206 may continuously rotate about the couch 216 when the couch 216 is inserted into the treatment area. In another example, the gantry 206 may rotate to a predetermined position when the couch 216 is inserted into the treatment area. For example, the gantry 206 may be configured to rotate the treatment output 204 about an axis ("A"). Both the couch 216 and the radiation treatment output 204 may be independently movable to other positions around the patient, such as being movable in a transverse direction ("T"), movable in a lateral direction ("L"), or rotating about one or more other axes, such as rotating about a transverse axis (indicated by "R"). A controller communicatively connected to one or more actuators (not shown) can control the movement or rotation of the couch 216 to properly position the patient in or out of the radiation beam 208 according to the radiation therapy plan. Both the couch 216 and the gantry 206 can move independently of each other in multiple degrees of freedom, which allows the patient to be positioned so that the radiation beam 208 can be precisely targeted to the tumor. The MLC can be integrated and included in the gantry 206 to deliver a radiation beam 208 of a specific shape.

[0074] Figure 2A The origin of the coordinate system shown in , including axis A, axis T, and axis L, can be located at the isocenter 210. The isocenter can be defined as the location where the central axis of the radiation beam 208 intersects the origin of the coordinate axes, such as to deliver a prescribed radiation dose to or within a patient. Alternatively, the isocenter 210 can be defined as the location where the central axis of the radiation beam 208 intersects the patient for various rotational positions of the radiation treatment output 204 positioned by the gantry 206 about axis A. As discussed herein, the gantry angle corresponds to the position of the gantry 206 relative to axis A, but the gantry angle can be determined with reference to any other axis or combination of axes.

[0075] The gantry 206 may also have an attached imaging detector 214. The imaging detector 214 is preferably positioned opposite the radiation source, and in an example, the imaging detector 214 may be located within the field of the radiation beam 208. The imaging detector 214 may be mounted on the gantry 206 (preferably opposite the radiotherapy output 204), for example, to maintain alignment with the radiation beam 208. The imaging detector 214 rotates about the rotation axis when the gantry 206 rotates. In one example, the imaging detector 214 may be a flat panel detector (e.g., a direct detector or a scintillator detector). In this manner, the imaging detector 214 may be used to monitor the radiation beam 208, or the imaging detector 214 may be used to image the patient's anatomical structure, such as portal imaging. The control circuitry of the radiotherapy device 202 may be integrated within the radiotherapy system 100 or remote from the radiotherapy system 100.

[0076] In the illustrative example, one or more of the couch 216, the treatment output 204, or the gantry 206 can be automatically positioned, and the treatment output 204 can establish the radiation beam 208 according to the specified dose for a particular treatment delivery instance. The order of treatment delivery can be specified according to the radiation treatment plan, such as using different orientations or positions of one or more of the gantry 206, the couch 216, or the treatment output 204. The treatment deliveries can be performed in sequence, but can intersect at a desired treatment site on or in the patient, such as at the isocenter 210. Thus, a prescribed cumulative dose of radiation therapy can be delivered to the treatment site while reducing or avoiding damage to tissue near the treatment site.

[0077] Figure 2B A radiation therapy device 202 is shown that may include a combination of a Linac and an imaging system (e.g., a CT imaging system). The radiation therapy device 202 may include an MLC (not shown). The CT imaging system may include an imaging X-ray source 218 that provides X-ray energy, for example, in the kiloelectronvolt (keV) energy range. The imaging X-ray source 218 may provide a fan-shaped and / or cone-shaped radiation beam 208 directed to an imaging detector 222 (e.g., a flat panel detector). The radiation therapy device 202 may be similar to the method described with respect to Figure 2A The depicted system, for example, includes a radiation therapy output 204, a gantry 206, a couch 216, and another imaging detector 214 (eg, a flat panel detector). An X-ray source 218 may provide relatively low energy diagnostic X-ray beams for imaging.

[0078] exist Figure 2BIn the illustrative example of FIG. 2 , the radiotherapy output 204 and the x-ray source 218 can be mounted on the same rotating gantry 206, rotated 90 degrees apart from each other. In another example, two or more x-ray sources can be mounted along the circumference of the gantry 206, for example, each x-ray source has its own detector arrangement to provide diagnostic imaging at multiple angles simultaneously. Similarly, multiple radiotherapy outputs 204 can be provided.

[0079] Figure 3 A radiation therapy system 300 is depicted, which may include a combination of a radiation therapy device 202 and an imaging system, such as a magnetic resonance (MR) imaging system (e.g., known in the art as MR-Linac) consistent with the disclosed examples. As shown, the system 300 may include a couch 216, an image acquisition device 320, and a radiation delivery device 330. The system 300 delivers radiation therapy to a patient according to a radiation therapy plan. In some examples, the image acquisition device 320 may correspond to Figure 1 The image acquisition device 170 in the image acquisition device can acquire an original image of a first modality (for example, an MRI image) or a destination image of a second modality (for example, a CT image).

[0080] The couch 216 can support a patient (not shown) during treatment. In some embodiments, the couch 216 can move along a horizontal translation axis (labeled "I") so that the couch 216 can move a patient placed on the couch 216 into and / or out of the system 300. The couch 216 can also rotate around a central vertical rotation axis that is transverse to the translation axis. To allow such movement or rotation, the couch 216 can have a motor (not shown) that enables the couch 216 to move in various directions and rotate around various axes. A controller (not shown) can control these movements or rotations so as to properly position the patient according to the treatment plan.

[0081] In some examples, the image acquisition device 320 may include an MRI machine for acquiring two-dimensional or three-dimensional MRI images of a patient before, during, and / or after a treatment process. The image acquisition device 320 may include a magnet 321 for generating a main magnetic field for magnetic resonance imaging. The magnetic field lines generated by the operation of the magnet 321 may extend substantially parallel to the central translation axis I. The magnet 321 may include one or more coils having axes extending parallel to the translation axis I. In some examples, one or more coils in the magnet 321 may be spaced apart so that a central window 323 of the magnet 321 is free of coils. In other examples, the coils in the magnet 321 may be sufficiently thin or have a reduced density so that they are substantially transparent to the wavelength radiation generated by the radiotherapy device 330. The image acquisition device 320 may also include one or more shielding coils that may generate a magnetic field of approximately equal magnitude and opposite polarity outside the magnet 321 to offset or reduce any magnetic field outside the magnet 321. As described below, the radiation source 331 of the radiation delivery device 330 may be positioned in an area where the magnetic field is canceled (cancelled by at least 1 order of magnitude, or reduced).

[0082] The image acquisition device 320 may also include two gradient coils 325 and 326, which may generate a gradient magnetic field superimposed on the main magnetic field. The gradient coils 325 and 326 may generate a gradient in the synthesized magnetic field, which allows the protons to be spatially encoded so that the position of the protons can be determined. The gradient coils 325 and 326 may be positioned around a common central axis with the magnet 321 and may be displaced along the central axis. The displacement may create a gap or window between the coils 325 and 326. In an example where the magnet 321 may also include a central window 323 between the coils, the two windows may be aligned with each other.

[0083] In some examples, the image acquisition device 320 may be an imaging device other than MRI, such as X-ray, CT, CBCT, spiral CT, PET, SPECT, optical tomography, fluorescence imaging, ultrasound imaging, radiotherapy viewport imaging device, etc. As recognized by those of ordinary skill in the art, the above description of the image acquisition device 320 relates to certain examples and is not intended to be limiting.

[0084] The radiation delivery device 330 may include a radiation source 331, such as an X-ray source or Linac, and an MLC 332. The radiation delivery device 330 may be mounted on a chassis 335. When the couch 216 is inserted into the treatment area, one or more chassis motors (not shown) may rotate the chassis 335 around the couch 216. In one example, when the couch 216 is inserted into the treatment area, the chassis 335 may continuously rotate around the couch 216. The chassis 335 may also have an attached radiation detector (not shown), preferably located opposite the radiation source 331, and the rotation axis of the chassis 335 is located between the radiation source 331 and the detector. In addition, the radiation delivery device 330 may include a control circuit (not shown) for controlling, for example, one or more of the couch 216, the image acquisition device 320, and the radiation therapy device 330. The control circuit of the radiation delivery device 330 may be integrated into the system 300 or remote from the system 300.

[0085] During a radiation therapy procedure, a patient may be positioned on couch 216. System 300 may then move couch 216 into a treatment area defined by magnet 321, coil 325, coil 326, and chassis 335. Control circuitry may then control radiation source 331, MLC 332, and chassis motors to deliver radiation to the patient through a window between coil 325 and coil 326 according to a radiation therapy plan.

[0086] Figure 2A , Figure 2B and Figure 3 An example of a radiation therapy device configured to provide radiation therapy to a patient using a configuration in which the radiation therapy output can rotate about a central axis (e.g., axis "A") is generally shown. Other radiation therapy output configurations can be used. For example, the radiation therapy output can be mounted to a robotic arm or manipulator with multiple degrees of freedom. In yet another example, the treatment output can be fixed, such as located in an area laterally separated from the patient, and a platform supporting the patient can be used to align the radiation therapy isocenter with a designated target in the patient's body.

[0087] Figure 4 A comparative example of a Leksell Gamma Knife radiotherapy device 430 is shown, which provides such radiotherapy via gamma radiation. As a brief overview of a Gamma Knife device, its radiation is emitted by a large number of fixed radiation sources and is focused toward a defined target or treatment volume via a collimator (i.e., a passage or channel for obtaining a beam with a defined cross-section). The dose of gamma radiation provided by each radiation source is insufficient to damage intervening tissue. However, tissue destruction occurs when the radiation beams from all or some of the radiation sources intersect or converge, resulting in radiation reaching a tissue-damaging level. The point of convergence is hereinafter referred to as the "isocenter", but may also be referred to as the "focus".

[0088] like Figure 4 As shown, during the radiotherapy process, the patient 402 may wear a coordinate frame 420 to keep the patient's body part (e.g., head) stable while undergoing surgery or radiotherapy. The coordinate frame 420 and the patient positioning system 422 may establish a spatial coordinate system that may be used when imaging the patient or during radiosurgery. The radiotherapy device 430 may include a protective housing 414 to surround a plurality of radiation sources 412. The radiation source 412 may generate a plurality of radiation beams (e.g., beamlets) through a beam channel 416. The plurality of radiation beams may be configured to focus on the isocenter 310 from different directions. Although each individual radiation beam may have a relatively low intensity, when multiple doses from different radiation beams converge at the isocenter 310, the isocenter 310 may receive a relatively high level of radiation. In some examples, the isocenter 310 may correspond to a target under surgery or treatment, such as a tumor.

[0089] Other types of radiotherapy devices (not shown) use protons and / or ions to deliver radiation therapy. The direction and shape of the radiation beam should be precisely controlled to ensure that the tumor receives the prescribed radiation dose and that the radiation from the beam minimizes damage to surrounding healthy tissue, particularly organs at risk (OARs). Therefore, treatment planning is used to design and control radiation beam parameters, and the radiotherapy device achieves treatment by delivering a spatially varying dose distribution to the patient.

[0090] Treatment plan optimization for radiotherapy (e.g., Gamma Knife radiosurgery) aims to maximize the dose delivered to a target volume in the patient (e.g., in tumor treatment) while minimizing the dose delivered to adjacent normal tissue. In treatment plan optimization, the delivered radiation dose is primarily constrained by two competing factors: the first is to deliver a high dose to the target volume, and the second is to deliver a low dose to surrounding normal tissue. The process of treatment plan optimization involves optimizing the number of shots used, the sector collimator combination, the duration of the shots, and the location of the shots relative to the patient's anatomy (i.e., the isocenter). Obviously, the irregularity and size of the target volume greatly influence the number of shots and shot sizes required for optimized treatment. Therefore, for Gamma Knife radiotherapy, the selected isocenter locations and their corresponding shots for a given case constitute the treatment plan.

[0091] The formal definition of the Pareto surface used in this paper is provided below: Let Ω be R n or an open subset of the n-dimensional surface, and let f 1 , f 2 , ..., f m :Ω→R∪{∞} is an extended real-valued function. Point It is called the Pareto optimal point if there does not exist such x∈Ω such that for all i=1...m And for some j there is f j (x)<f j (x). i , for all of The collection of is called the Pareto surface.

[0092] As provided in this article, the optimized form is:

[0093]

[0094] The set of solutions x* therefore depends on the weight vector w, i.e. x*=x*(w), and the i-th objective function value of the solution is f i (x(w))≡f i (w). For this type of optimization problem, it can be seen that for any weight w, the solution x is a non-dominated point Furthermore, the dimension of the Pareto surface is n-1, since the solution to Equation 3 is independent of scaling the weights by a scalar value (i.e., (w 1 , w 2 , ..., w n ) and λ(w 1 , w 2 , ..., w n ) give the same solution).

[0095] The generalized surface that is the object of this optimization is called a "differentiable manifold". The surface locally looks like R n In vector spaces, we can define everything from open sets on manifolds, graphs to R n Mapping of open sets in The entire manifold is covered with a set of graphs, called an atlas. The manifold locally looks like R n The implication of suggests the possibility of performing differentiable computations on manifolds. For example, if is an open set U on M (with n Mapping ), and there is a second (V, ψ) open set V on M (with n ψ), and there is a partial overlap between U and V (i.e. ),but is necessarily differentiable. A simple example is a sphere, which can be covered with at least two graphs (e.g., one that excludes the North Pole and another that excludes the South Pole). Note that even though the sphere is embedded in three dimensions, you don't need to know this to describe its two-dimensional surface geometrically. In fact, there are some examples of two-dimensional surfaces that cannot be embedded in two dimensions, such as the Klein bottle, but can still be described using a finite number of two-dimensional graphs.

[0096] Depending on the nature of the optimization problem, the Pareto surface can be quite complex, with several branches that can cross each other and become disconnected. However, assuming second-order differentiability of the objective function, the Pareto surface is a part of an (m-1)-dimensional manifold. In the case of a convex objective function, the Pareto surface is simply a convex (m-1)-dimensional manifold that can be covered by several (possibly one) graphs.

[0097] In one example, a regular submanifold of a k-dimensional manifold is one that locally looks like R k A submanifold S is a subset of a manifold M. A submanifold inherits the differentiable structure of a manifold. A circle is a submanifold of a sphere (one of an infinite number of submanifolds), which is in turn a 3D sphere x 2 +y 2 +z 2 +u 2 =R 2 A submanifold of n-2, n-3, ... 1 dimensions is a submanifold of n-2, n-3, ... 1 dimensions. By changing only one of the weights, the submanifold is a line along which the objective function varies. We choose a fixed weight, called We have a single point on the Pareto surface P, namely p s By changing the weight of each factor at a time, We obtain n-1 lines on the Pareto surface, all of which pass through p s These lines form a grid or coordinate system on the Pareto surface, close to p s . Very close to p s When , these lines become tangent vectors to the surface. s The set of all tangent vectors at is called the tangent space and is represented as The union of all tangent spaces on a manifold is called the tangent bundle Note that in general if p s ≠q s but

[0098] Accordingly, reference to navigation or "surfing" as used herein refers to navigating from a particular point p on P s We start with a cost function and then explore how the cost function value changes as we travel in a given direction. Due to the high dimensionality of the Pareto surface, it is computationally more tractable than the entire Pareto surface and it is easier to visualize the changes in the subpopulation. In addition, the direction of movement may be unknown and becomes part of the navigation. For example, it may be beneficial to minimize one objective while controlling for the deterioration of other objectives. Therefore, at each point p s Found everywhere The tangent vector in minimizes the objective. By moving infinitely along the tangent vector, nearby points can be reached. This process is repeated until the user decides that other objectives are too bad. As described below, this simple navigation can be expressed as an ordinary differential equation, and the solution is a path on the Pareto surface.

[0099] Figure 5 A high-level view of the operation of the radiation therapy planning workflow is provided. Specifically, this workflow includes the use of Pareto navigation and exploration as part of the radiation therapy problem optimization process, which evaluates and formulates a Pareto optimal solution to the radiation therapy problem.

[0100] Figure 5 5 shows in more detail how radiation therapy question information 510 for treatment of a human subject is provided with information such as a target region for treatment 512 and a region at risk 514 of an organ (or a defined region of healthy tissue). Other information related to the radiation therapy question may include radiation therapy machine information 520, such as machine capabilities 522. Thus, the parameters (criteria) of the radiation therapy question may correspond to clinical preferences, anatomical regions, detailed definitions of regions to receive (or not receive) treatment, etc.

[0101] Given these parameters and information, the process of solving the radiation therapy problem 530 is performed to generate a plurality of radiation therapy problem solutions 540. These solutions may be generated using a variety of problem solvers to generate an initial set of Pareto optimal radiation therapy solutions. As described above, the set of Pareto optimal radiation therapy solutions may constitute or define a Pareto surface, which is explored using the following navigation techniques.

[0102] Pareto optimal solution navigation 550 may include the following functions or capabilities:

[0103] Directed Exploration 552, which enables transformation, depiction, and navigation of solutions within a representation of the solution space;

[0104] A selection function 554 that enables the selection and identification of specific parameter weights (i.e., parameter values) within the solution space;

[0105] A restriction function 556 that enables exclusion of certain parameter weights (ie, parameter values) within the solution space; and

[0106] An identification function 558 is capable of identifying a new selected solution (ie, an additional solution) at a particular location within the solution space.

[0107] The new selected solution provides a particular combination of weight values ​​that can be used to generate radiation treatment data and ultimately generate a particular radiation treatment plan in operation 560. The generation of the radiation treatment plan (and the associated selection or modification of the radiation treatment solution) in operation 560 may depend on machine capabilities 522. Finally, the generated treatment plan (and data associated with the treatment plan) may be used to deliver radiation treatment in operation 570.

[0108] In one example, navigation and directed exploration of Pareto optimal solutions are performed using submanifolds. In one example, the submanifold is one-dimensional, meaning that the coordinate graph is a one-dimensional path in the weight space and the resulting solution to the MCO problem belongs to the zero-order set of optimality conditions. Navigation then reduces to solving ordinary differential equations (ODEs) based on directed derivatives. There are some sophisticated ODE solvers that can, for example, track the accuracy of the solution and adaptively determine when functions (and possibly derivatives) need to be evaluated to achieve a certain level of accuracy. In this setting, such a query may be equivalent to solving an optimization problem, and given that the ODE solver may be able to produce an accurate initial guess, a warm start method may be preferably used to complete the solution.

[0109] As you can see, there are three different but related spaces involved in exploring solutions:

[0110] i) The vector of objective function values, f(x) = (f 1 (x), …, f m (x)), exists in the target space, which is usually R m (m is the number of targets).

[0111] ii) The weights w exist in the weight space. Technically, the multi-criteria optimization problem is a vector optimization problem that seeks to minimize the vector of the objective function with respect to some cone K, and the weight space is the set of non-negative vectors in the dual cone K*.

[0112] iii) The decision variables x (interchangeably referred to as optimization variables in this paper) exist in the decision variable space, which is an n-dimensional space (usually R n Here, n>m, but due to the direct correspondence between m-dimensional weights and solutions, the set of Pareto optimal solutions (in the decision variable space) is an m-1-dimensional manifold embedded in the higher-dimensional decision variable space.

[0113] Therefore, the “solution” used in this paper can refer to the vector of decision variables x*, or the objective function value f(x * ), or both. Similarly, the concept of a Pareto surface may refer to the set of Pareto optimal solutions in the objective space, the decision variable space, or both.

[0114] Figure 6 An example workflow for identifying and exploring additional Pareto optimal solutions is provided. In one example, the workflow can be implemented as part of a radiation therapy planning and optimization process by navigating 550 (and sub-functions 552, 554, 556, 558) of Pareto optimal solutions. However, the workflow can also be implemented as a stand-alone evaluation process, or other aspects of solution planning and exploration.

[0115] The workflow begins at operation 610 by obtaining a Pareto surface (or a sufficiently large set) of potential solutions to the radiation therapy problem. This solution space provides multiple solutions to the radiation therapy problem (a multi-criteria optimization problem) that can be adjusted by weights for multiple criteria. Solutions to the radiation therapy problem provide a subset of the Pareto surface (in principle, each Pareto optimal point can be found by solving the radiation therapy problem for each choice of weights).

[0116] In operation 620, a set of path-connected weights is defined within a submanifold of the Pareto surface. The dimension of the submanifold space is less than the cardinality of the multiple weights used by the standard for the radiation therapy problem. The cardinality used herein can be understood by the following example: if there are N solutions (e.g., hundreds of solutions), each solution corresponds to M weights (e.g., a dozen or fewer weights), then the dimension of the submanifold is less than M (at most it can be M-1, because the solutions are invariant with respect to the scaling of the objective function).

[0117] In one example, the submanifold space provides a one-dimensional path for navigation on the path. However, in another example, a multi-dimensional surface can also be used to navigate the submanifold space. Consider a Pareto surface with N solutions (e.g., hundreds of solutions), each corresponding to M weights (e.g., a dozen or fewer weights). In this case, the dimension of the submanifold is less than M (it can be at most M-1, because the solutions are invariant with respect to the scaling of the objective function).

[0118] In operation 630, navigation in the submanifold is simplified to an ordinary differential equation (ODE) based on the parameterized direction of the path. In one example, the differential equation is formulated on the set of weights connected by the path based on a first-order optimality condition. For example, navigation of a one-dimensional path can start from a Pareto optimal initial point because navigation is simplified to an ordinary differential equation based on directional derivatives. For constrained optimization problems, meeting the Karush-Kuhn-Tucker (KKT) condition is a first-order optimality condition.

[0119] In one example, exploration is provided by solving a boundary value problem, wherein an initial point and a final point are provided for navigation of a one-dimensional path. In order to solve the boundary value problem, navigation interpolates between the initial point and the final point. In another example, exploration is provided by solving an initial value problem, wherein an initial point is provided for navigation of a one-dimensional path. In this example, navigation operates until a predetermined stop condition is satisfied. The predetermined stop condition can be: a maximum acceptable degradation of a time, distance, or clinical indicator. For example, the maximum acceptable degradation of a clinical indicator may also involve using a simultaneous localization and mapping (SLAM) method to improve a first clinical indicator while minimizing degradation of a second clinical indicator.

[0120] In some examples, the radiation therapy problem is a constrained problem and at least a portion of the manifold space has a non-differentiable portion. (This occurs when the Pareto surface may not be differentiable, that is, when the active set (i.e., the set of equal constraints) changes. Because optional operations 640 and 650 address these situations. In operation 640, non-differentiable regions are identified. In operation 650, constraints are removed from these non-differentiable regions. In one example, the non-differentiable portion of the manifold space is smoothly approximated using a least squares method. In another example, a barrier formulation is used to convert the constrained problem into an unconstrained problem by including the constraint as a term in the objective function, associating a penalty for violating the constraint.

[0121] In operation 660, differentiation is performed by solving a differential equation (ODE) to generate additional Pareto optimal solutions. In one example, such differentiation can be performed by automatic differentiation techniques. In another example, such differentiation can be performed analytically, such as evaluating derivatives from vector-Jacobian products (thus, avoiding the need to calculate and store a full Jacobian matrix).

[0122] In operation 670, the newly generated (additional) Pareto optimal solution is explored. Here, such exploration can be performed using a pre-provided or dynamically provided coordinate chart, which is similar to the exploration determined by open-loop or closed-loop control, respectively. As described above, such exploration is controlled based on the solution of the boundary value problem or the initial value problem and the accompanying stopping condition.

[0123] In evaluation 680, a determination is made as to whether a satisfactory solution has been found from the additional Pareto optimal solutions. The determination may be made based on evaluation criteria or user input. If a satisfactory solution has been found, then the solution is used in the radiation therapy plan in operation 690. If a satisfactory solution has not been found, then additional exploration and path finding operations are repeatedly performed (e.g., operations 620-670 are repeated).

[0124] As mentioned above, one may encounter non-differentiable parts of the Pareto surface, which are solved by smooth approximations and barrier constraints. The constraints will restrict the set of feasible solutions compared to the unconstrained problem, thus changing the Pareto surface. For some weight combinations, the "unconstrained solution" will still be feasible with the constraints, but for some other combinations the Pareto surface will be modified. On the Pareto surface, there is no smooth connection in the tangent space from the unchanged part of the Pareto surface to the part of the Pareto surface modified by the constraints, which means that navigation methods based on solving ordinary or partial differential equations will fail. Near "ridges", one can modify the Pareto surface so that it varies smoothly, so that the tangent spaces connect smoothly, which means that the navigation methods described above will apply.

[0125] Another way to formulate an optimization problem with constraints is to introduce a barrier function into the objective function. The idea is that the closer the solution is to violating the constraint, the higher the value of the barrier function. If the constraint is of the form Ax≤b, then a typical barrier function is:

[0126]

[0127] High costs close to constraints (and infinite costs both inside and outside them) lead to solutions that do not violate the constraints. Since optimization problems consist of smooth and differentiable terms, replacing constraints with barrier penalty terms leads to a smooth and differentiable Pareto surface. The barrier function achieves the same functionality as the smooth approximation, but without explicitly introducing terms to smooth the Pareto surface.

[0128] Although not directly depicted in the figures discussed herein, various forms of user interfaces or representations may be provided to enable representation of the solution space for a radiation therapy problem based on multiple Pareto optimal solutions. This may be provided in a graphical user interface having functionality for configuring a treatment plan and receiving and outputting data related to the treatment plan. For example, information related to a solution to a radiation therapy problem for a particular set of parameters may be displayed. User interaction may be obtained in such a user interface to modify a particular set of parameters. It will be appreciated that various interactive Pareto navigation functions may be provided in the user interface, such as an approximation of a Pareto surface that lets a user explore estimates of solutions corresponding to new or different parameter values ​​that have not been previously evaluated.

[0129] It will be appreciated that the Pareto optimal plans generated using such techniques can be viewed as points on a Pareto surface. The more points generated, the better the description of the surface. As the user explores the Pareto surface, the user will be doing some type of interpolation between the existing calculation points in order to select and use a solution. It is likely that the plan that the user will eventually develop and select will not necessarily be the plan of one of the calculation points.

[0130] In some examples, all Pareto optimal solutions, or at least a substantial subset of Pareto optimal solutions, may be generated and evaluated before a user selects a particular plan or solution. Thus, the plan ultimately selected may be a plan that is "between" two generated points on the Pareto surface. Using the techniques discussed herein, Pareto surface navigation may be performed quickly and efficiently.

[0131] Figure 7 A representation 700 is shown that is explored in a Pareto surface. Here, three dimensions are shown: a first axis 710 representing the value of the total irradiation time for the treatment (e.g., a measure of the beam on time (BOT) for a particular radiotherapy solution); a second axis 720 representing the value of selectivity (i.e., the fraction of the volume of dose that is equal to or greater than the planned treatment dose constrained by the target volume, defining the selectivity between the inner and outer regions of the region of interest for a particular radiotherapy solution); and a third axis 730 representing the coverage of the radiotherapy solution (i.e., the fraction of the target volume that receives a dose equal to or greater than the planned dose, defining the coverage of the treatment region for a particular radiotherapy solution). Three-dimensional modeling and rendering of the representation 700 allows exploration of the Pareto surface along a one-dimensional path (path 740).

[0132] Here, exploration along path 740 demonstrates that a treatment plan with a 25% reduction in treatment time (e.g., beam on time) can be identified by accepting a small decrease (e.g., at most 2%) in quality (e.g., coverage and / or selectivity). This path follows the steepest negative gradient in the treatment time dimension, satisfying the maximum quality decrease allowed.

[0133] FIG. 8A to FIG. 8D A further representation explored in Pareto surfaces is provided. First, Fig. 8A The set of Pareto optimal points for an optimization case with four objectives is shown, with the Pareto surface rendered in 3D. The size of the shaded (dark in the shade) and filled circles is proportional to the value of the fourth objective term.

[0134] Figure 8B Shows the Fig. 8A The result of smoothing a portion of the Pareto surface. As mentioned above, smoothing provides a mechanism for efficient Pareto surfing.

[0135] Figure 8C The result of fixing the fourth objective to a specific value to produce a 2D submanifold is shown. In this figure, interpolation has been applied to show the surface.

[0136] Fig.8D The results of fixing multiple targets (f3 and f4) to specific values ​​are shown. Here, this results in a representation of a 1D submanifold, representing a 1D path on a 3D Pareto surface.

[0137] Fig. 9A flowchart 900 of a method for radiation therapy planning based on the above techniques is shown. For example, the following features of flowchart 900 may be used in conjunction with reference to Figure 5 The navigation and optimization operations discussed, reference Figure 6 Solution exploration operations discussed and references Figure 7 and Figure 8A-8D The path exploration operation in question is integrated or adapted.

[0138] Operation 910 begins with the operation of obtaining a radiation therapy problem that defines various parameters of radiation therapy delivery from a radiation therapy machine. Such a radiation therapy problem may be adjusted by parameters that are optionally received with a request to solve the radiation therapy problem.

[0139] Operation 920 continues with operations to generate solutions for the radiation therapy problem, which may be performed for a single problem until a solution for the entire Pareto surface or frontier (or a sufficient portion of the Pareto surface or frontier) is identified. Although not shown, the treatment plan optimization operation may be based on dose, geometry, imaging, machine learning, radiobiology, or other relevant factors. A variety of solution optimizers and solution methods may be used to generate usable solutions.

[0140] Operation 930 continues the process of identifying additional radiotherapy solutions by exploring Pareto optimal radiotherapy solutions. This can be performed in sub-operations, including: in operation 932, establishing a submanifold space from the solutions in the Pareto surface; in operation 934, generating additional weights in the submanifold space based on the derivative of the optimality condition; in operation 936, performing navigation along the path in the submanifold space to obtain additional solutions; and in operation 938, selecting a weight set from the additional solutions.

[0141] In operation 940, treatment planning data is generated using the additional solutions, and the additional solutions are further optimized, if applicable, and implemented as part of the treatment plan. For example, the additional solutions may provide a warm start to identify a solution for treatment planning, receive additional optimizations, and implement the additional optimizations to create a treatment plan. Consistent with the above example, treatment planning data for radiation therapy may include a set of treatment delivery parameters corresponding to the capabilities and operational controls of the radiation therapy machine.

[0142] Fig.10A block diagram of an example of a machine 1000 is shown on which one or more of the methods discussed herein may be implemented. In one or more examples, one or more items of the radiation therapy processing computing system 110 may be implemented by the machine 1000. In alternative embodiments, the machine 1000 operates as a standalone device, or may be connected (e.g., networked) to other machines. In one or more examples, the radiation therapy processing computing system 110 may include one or more of the items of the machine 1000. In a networked deployment, the machine 1000 may operate in the capacity of a server or client machine in a server-client network environment, or as a peer machine in a peer-to-peer (or distributed) network environment. The machine may be a personal computer (PC), a server, a tablet, a smartphone, a network appliance, an edge computing device, a network router, a switch or a bridge, or any machine capable of executing instructions (sequentially or otherwise) that specify actions to be taken by the machine. Further, while a single machine is illustrated, the term "machine" shall also be taken to include any collection of machines that individually or jointly execute a set (or multiple sets) of instructions to perform any one or more of the methodologies discussed herein.

[0143] The example machine 1000 includes a processing circuit or processor 1002 (e.g., a CPU, a graphics processing unit (GPU), an ASIC, a circuit system (e.g., one or more transistors, resistors, capacitors, inductors, diodes, logic gates, multiplexers, buffers, modulators, demodulators, radios (e.g., transmitting or receiving radios or transceivers)), a sensor 1021 (e.g., a transducer that converts one form of energy (e.g., light, heat, electricity, mechanical or other energy) into another form of energy), or similar devices or combinations thereof), a main memory 1004, and a static memory 1006, which are in communication with each other via a bus 1008. The machine 1000 (e.g., a computer system) may also include a video display device 1010 (e.g., a liquid crystal display (LCD) or a cathode ray tube (CRT)). The machine 1000 also includes an alphanumeric input device 1012 (e.g., a keyboard), a user interface (UI) navigation device 1014 (e.g., a mouse), a disk drive or mass storage unit 1016, a signal generating device 1018 (e.g., a speaker), and a network interface device 1020.

[0144] The disk drive unit 1016 includes a machine-readable medium 1022 on which is stored one or more sets of instructions and data structures (e.g., software) 1024 that embody or are utilized by any one or more of the methods or functions described herein. The instructions 1024 may also reside, in whole or in part, within the main memory 1004 and / or within the processor 1002 during execution of the instructions 1024 by the machine 1000, the main memory 1004 and the processor 1002 also constituting machine-readable media.

[0145] As shown, machine 1000 includes output controller 1029. Output controller 1029 manages the flow of data into and out of machine 1000. Output controller 1028 is sometimes referred to as a device controller, where software that interacts directly with output controller 1028 is called a device driver.

[0146] Although the machine-readable medium 1022 is shown as a single medium in one example, the term "machine-readable medium" may include a single medium or multiple media (e.g., centralized or distributed databases, and / or associated caches and servers) that store one or more instructions or data structures. The term "machine-readable medium" should also be considered to include any tangible medium that can store, encode, or carry instructions or can store, encode, or carry data structures that are executed by a machine and cause the machine to perform any one or more of the methods of the present disclosure, which data structures are utilized by or associated with such instructions. Therefore, the term "machine-readable medium" should be considered to include, but are not limited to, solid-state memory and optical and magnetic media. Specific examples of machine-readable media include non-volatile memory, including by way of example: semiconductor memory devices, such as erasable programmable read-only memory (EPROM), EEPROM, and flash memory devices; magnetic disks, such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks.

[0147] The instructions 1024 may also be transmitted or received over a communication network 1026 using a transmission medium. The instructions 1024 may be transmitted using a network interface device 1020 and any of a number of well-known transmission protocols (e.g., HTTP). Examples of communication networks include LANs, WANs, the Internet, mobile phone networks, plain old telephone (POTS) networks, and wireless data networks (e.g., WiFi and 4G / 5G data networks). The term "transmission medium" shall be deemed to include any intangible medium that is capable of storing, encoding, or carrying instructions to be executed by a machine, and includes digital or analog communication signals or other intangible media to facilitate communication of such software.

[0148] As used herein, "communicatively coupled between" means that any of the coupled entities must communicate through an item therebetween, and the entities cannot communicate with each other without communicating through the item.

[0149] Additional Notes

[0150] The above detailed description includes reference to the accompanying drawings, which form a part of the detailed description. The accompanying drawings show specific embodiments in which the present disclosure can be practiced by way of illustration and not by way of limitation. These embodiments are also referred to as "examples" in this article. These examples may include elements other than those elements shown or described. However, the inventors also contemplate providing examples of only those elements shown or described. In addition, whether relative to a specific example (or one or more aspects thereof) or relative to other examples (or one or more aspects thereof) shown or described in this article, the inventors also contemplate examples of any combination or permutation (or one or more aspects thereof) of those elements shown or described.

[0151] All publications, patents, and patent documents mentioned in this document are incorporated herein by reference in their entirety, as if individually incorporated by reference. If there are inconsistent usages between this document and those documents incorporated by reference, the usage in the incorporated reference(s) should be considered supplementary to that of this document; for conflicting inconsistencies, the usage in this document controls.

[0152] In this document, the terms "a", "an", "the", and "said" are used when introducing elements in various aspects of the present disclosure or embodiments thereof, as is common in patent documents, to include more than one or more of the elements, regardless of any other instances or uses of "at least one" or "one or more". In this document, the term "or" is used to indicate a non-exclusive, or, such that, unless otherwise specified, "A or B" includes "A but not B", includes "B but not A", and includes "A and B".

[0153] In the following claims, the terms "including" and "in which" are used as the plain-English equivalents of the respective terms "comprising" and "wherein." Furthermore, in the following claims, the terms "comprising," "including," and "having" are intended to be open-ended, meaning that there may be additional elements in addition to the listed elements, such that anything following such a term in a claim (e.g., comprising, including, having) is still considered to fall within the scope of the claim. Moreover, in the following claims, the terms "first," "second," and "third," etc. are used merely as labels, and are not intended to impose numerical requirements on their objects.

[0154] Embodiments of the present disclosure can be implemented with computer executable instructions.Computer executable instructions (such as software code) can be organized into one or more computer executable components or modules.Any number of such components or modules and any organization of such components or modules can be used to implement various aspects of the present disclosure.For example, various aspects of the present disclosure are not limited to the specific computer executable instructions or specific components or modules shown in the drawings and described in this article.Other embodiments of the present disclosure may include different computer executable instructions or components with more or less functions than those shown and described herein.

[0155] The method examples (e.g., operations and functions) described herein may be at least partially machine or computer implemented (e.g., implemented as software code or instructions). Some examples may include a computer-readable medium or machine-readable medium encoded with instructions, which are operable to configure an electronic device to perform the method as in the above example. The implementation of such a method may include software code such as microcode, assembly language code, high-level language code, etc. (e.g., "source code"). Such software code may include computer-readable instructions (e.g., "objects" or "executable code") for performing various methods. The software code may form part of a computer program product. The software implementation of the embodiments described herein may be provided via an article having codes or instructions stored thereon, or via a method of operating a communication interface to send data via a communication interface (e.g., wirelessly, via the Internet, via satellite communications, etc.).

[0156] In addition, the software code may be tangibly stored on one or more volatile or non-volatile computer-readable storage media during execution or at other times. These computer-readable storage media may include any mechanism for storing information in a form that can be accessed by a machine (e.g., a computing device, an electronic system, etc.), such as, but not limited to, a floppy disk, a hard disk, a removable disk, any form of disk storage media, a CD-ROM, a magneto-optical disk, a removable optical disk (e.g., a compact disk and a digital video disk), a flash memory device, a cassette, a memory card or stick (e.g., a secure digital card), a RAM (e.g., a CMOS RAM, etc.), a recordable / non-recordable medium (e.g., a read-only memory (ROM)), an EPROM, an EEPROM, or any type of medium suitable for storing electronic instructions, etc. Such a computer-readable storage medium is coupled to a computer system bus so that it can be accessed by the processor and other parts of the OIS.

[0157] In one embodiment, the computer readable storage medium may have an encoded data structure for treatment planning, wherein the treatment plan may be adaptive. The data structure for the computer readable storage medium may be at least one of a Digital Imaging and Communications in Medicine (DICOM) format, an extended DICOM format, an XML format, and the like. DICOM is an international communication standard that defines a format for transmitting data related to medical images between various types of medical devices. DICOMRT refers to a communication standard specific to radiation therapy.

[0158] In various embodiments of the present disclosure, the methods of creating components or modules can be implemented in software, hardware, or a combination thereof. For example, the methods provided by various embodiments of the present disclosure can be implemented in software using standard programming languages ​​such as, for example, C, C++, C#, Java, Python, CUDA, etc., and combinations thereof. As used herein, the terms "software" and "firmware" are interchangeable and include any computer program stored in a memory for execution by a computer.

[0159] The communication interface includes any mechanism that interfaces with any of hardwired media, wireless media, optical media, etc. to communicate with another device, such as a memory bus interface, a processor bus interface, an Internet connection, a disk controller, etc. The communication interface can be configured by providing configuration parameters and / or sending signals to prepare the communication interface to provide data signals describing the software content. The communication interface can be accessed via one or more commands or signals sent to the communication interface.

[0160] The present disclosure also relates to a system for performing the operations herein. The system can be constructed specifically for the desired purpose, or can include a general-purpose computer selectively activated or reconfigured by a computer program stored in a computer. Unless otherwise stated, the operation or execution order of the operations in the embodiments of the present disclosure shown and described herein is not necessary. That is, unless otherwise stated, operations can be performed in any order, and compared with those operations disclosed herein, the embodiments of the present disclosure can include other operations or fewer operations. For example, it is contemplated that running or performing a specific operation before another operation, simultaneously with another operation, or after another operation falls within the scope of the various aspects of the present disclosure.

[0161] In view of the above, several objects of the present disclosure can be achieved and other advantageous results are obtained. Having described various aspects of the present disclosure in detail, it will be apparent that modifications and variations are feasible without departing from the scope of various aspects of the present disclosure as defined in the appended claims. Since various changes can be made in the above constructions, products and methods without departing from the scope of various aspects of the present disclosure, it is intended that all content contained in the above description and shown in the accompanying drawings should be interpreted as illustrative, and not in a restrictive sense.

[0162] The above description is intended to be illustrative, rather than restrictive. For example, the above examples (or one or more aspects thereof) can be used in combination with each other. In addition, without departing from the scope of the present disclosure, many modifications can be made to adapt specific situations or materials to the teachings of the present disclosure. Although the size and type of the materials and coatings described herein are intended to limit the parameters of the present disclosure, they are by no means restrictive, but exemplary embodiments. After reviewing the above description, many other embodiments are obvious to those skilled in the art. Therefore, the scope of the present disclosure should be determined with reference to the full scope of the equivalents to which the attached claims and such claims are assigned.

[0163] In addition, in the above detailed description, various features can be combined together to streamline the present disclosure. This should not be interpreted as an intention that unclaimed disclosed features are necessary for any claim. Instead, the subject matter of the invention may lie in less than all the features of a particular disclosed embodiment. Therefore, the following claims are hereby incorporated into the detailed description, with each claim existing independently as a separate embodiment. The scope of the present disclosure should be determined with reference to the appended claims and the full scope of equivalents to which such claims are assigned. In addition, the limitations of the appended claims are not written in a means-plus-function format, and are not intended to be interpreted based on the sixth paragraph of 35 U.S.C. §112, unless and until the limitation of such a claim explicitly uses the phrase "means for..." and thereafter describes the function without further structure.

[0164] The Abstract is provided to comply with 37 CFR §1.72(b), to allow the reader to quickly ascertain the nature of the technical disclosure. It is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims.

Claims

1. A computer-implemented method for radiation therapy planning, the method comprising: obtaining a plurality of solutions to a radiation therapy problem defined in a manifold space for providing radiation therapy, wherein the radiation therapy problem is a multi-criteria optimization problem, and wherein each of the plurality of solutions has a plurality of weights for adjusting a plurality of criteria of the multi-criteria optimization problem; exploring the plurality of solutions to the multi-criteria optimization problem to identify additional solutions in a submanifold space, the exploring comprising: establishing the submanifold space from the manifold space, the submanifold space representing the plurality of solutions in fewer dimensions than the plurality of weights; generating an additional set of weights in the submanifold space based on a derivative of a first-order optimality condition of the radiation therapy problem, the derivative being determined with respect to the plurality of weights; and performing navigation in the submanifold space to obtain the additional solution, the additional solution corresponding to one of the additional weight sets; and Treatment planning parameters are generated based on the additional solutions, wherein the treatment planning parameters are used in treatment planning for delivering the radiation therapy via a radiation therapy machine.

2. The method according to claim 1, wherein: The navigation is performed in the submanifold space using a one-dimensional path.

3. The method according to claim 2, wherein: The navigation of the one-dimensional path is performed starting from a Pareto optimal initial point, and wherein the navigation is reduced to finding the one-dimensional path by solving ordinary differential equations based on directional derivatives.

4. The method according to claim 2, wherein: In order to identify the additional solution a boundary value problem is solved, wherein an initial point and a final point are provided for the navigation of the one-dimensional path, and wherein the navigation interpolates between the initial point and the final point.

5. The method according to claim 2, wherein: To identify said additional solutions an initial value problem is solved, wherein an initial point is provided for said navigation of said one-dimensional path, and wherein said navigation operates until a predetermined stopping condition is satisfied.

6. The method according to claim 5, wherein: The predetermined stopping condition is: maximum acceptable degradation of time, distance or clinical index.

7. The method according to claim 6, wherein: The predetermined stopping condition is a maximum acceptable degradation of a clinical indicator, wherein a simultaneous localization and mapping (SLAM) method is used to perform the navigation on the one-dimensional path, wherein the SLAM method identifies an improvement in a first clinical indicator while minimizing a degradation in a second clinical indicator.

8. The method according to claim 1, wherein: The set of additional weights in the submanifold space corresponds to a level set of the optimality condition for the radiation therapy problem.

9. The method according to claim 1, wherein: The derivatives are provided by automatic differentiation.

10. The method according to claim 1, wherein: At least a portion of the manifold space has a non-differentiable portion.

11. The method according to claim 10, wherein: The non-differentiable portion of the manifold space is smoothly approximated using a least squares method.

12. The method according to claim 10, wherein: The barrier formulation is used to convert a constrained problem into an unconstrained problem by including the constraints as terms in the objective function and associating a penalty for violating the constraints.

13. The method according to claim 1, wherein: The plurality of criteria for the radiation therapy question corresponds to clinical preferences, and wherein at least one criterion of the plurality of criteria relates to a specific anatomical region for receiving the radiation therapy from the radiation therapy machine.

14. The method according to claim 1, further comprising: A representation of a solution space is generated based on the plurality of solutions to the radiation therapy problem, wherein the solution space is a Pareto surface including a set of Pareto optimal solutions.

15. The method according to claim 1, further comprising: generating a display of a graphical user interface configured to provide functionality for configuring the treatment plan; and Information associated with the additional solutions is displayed in the graphical user interface.

16. The method of claim 1, further comprising: receiving a selection of the additional solution; Wherein the treatment planning parameters are generated based on the selection of the additional solution.

17. The method according to claim 16, wherein: The additional solutions provide a warm start to identify a solution for the treatment plan, the method further comprising: receiving an optimization of the additional solution; Wherein the treatment plan parameters are generated based on the optimization of the additional solution.

18. The method according to claim 1, wherein: The treatment planning parameters for the radiation therapy include a set of treatment delivery parameters corresponding to the capabilities of the radiation therapy machine.

19. The method according to claim 18, wherein: The treatment plan is for providing the radiation therapy using a gamma knife, and wherein the set of treatment delivery parameters includes a set of isocenters for delivering the radiation therapy.

20. The method according to claim 18, wherein: The set of treatment delivery parameters also includes a timing for delivering the radiation therapy, and a collimator sequence for delivering the radiation therapy.

21. The method according to claim 18, wherein: The treatment plan is used to deliver the radiation therapy using a linear accelerator or a magnetic resonance (MR)-linear accelerator radiation therapy machine.

22. The method according to claim 21, wherein: The treatment plan is for providing volumetric arc modulated therapy (VMAT) or intensity modulated radiation therapy (IMRT) to the radiation therapy, and wherein the set of treatment delivery parameters includes: a set of arc control points for one or more arcs, a fluence field, a gantry speed, and a dose rate along the one or more arcs.

23. A non-transitory computer-readable storage medium comprising computer-readable instructions for radiation therapy planning, wherein when the instructions are executed by a computing machine, the computing machine is caused to perform the method of any one of claims 1 to 22.

24. A computing system configured for radiation therapy planning, the computing system comprising: one or more memory devices for storing data of a radiation therapy problem for use in providing radiation therapy to a human subject from a radiation therapy machine; and One or more processors configured to: obtaining a plurality of solutions to the radiation therapy problem defined in a manifold space, wherein the radiation therapy problem is a multi-criteria optimization problem, and wherein each of the plurality of solutions has a plurality of weights for adjusting a plurality of criteria of the multi-criteria optimization problem; The plurality of solutions to the multi-criteria optimization problem are explored to identify additional solutions in a submanifold space by: establishing the submanifold space from the manifold space, the submanifold space representing the plurality of solutions in fewer dimensions than the plurality of weights; generating an additional set of weights in the submanifold space based on a derivative of a first-order optimality condition of the radiation therapy problem, the derivative being determined with respect to the plurality of weights; and performing navigation in the submanifold space to obtain the additional solution, the additional solution corresponding to one of the additional weight sets; and Treatment planning parameters are generated based on the additional solutions, wherein the treatment planning parameters are used in treatment planning for delivering the radiation therapy via a radiation therapy machine.

25. The computing system of claim 24, wherein: The navigation is performed in the submanifold space using a one-dimensional path.

26. The computing system of claim 25, wherein: The navigation of the one-dimensional path is performed starting from a Pareto optimal initial point, and wherein the navigation is reduced to finding the one-dimensional path by solving ordinary differential equations based on directional derivatives.

27. The computing system of claim 25, wherein: In order to identify the additional solution a boundary value problem is solved, wherein an initial point and a final point are provided for the navigation of the one-dimensional path, and wherein the navigation interpolates between the initial point and the final point.

28. The computing system of claim 25, wherein: To identify said additional solutions an initial value problem is solved, wherein an initial point is provided for said navigation of said one-dimensional path, and wherein said navigation operates until a predetermined stopping condition is satisfied.

29. The computing system of claim 28, wherein: The predetermined stopping condition is: maximum acceptable degradation of time, distance or clinical index.

30. The computing system of claim 29, wherein: The predetermined stopping condition is a maximum acceptable degradation of a clinical indicator, wherein a simultaneous localization and mapping (SLAM) method is used to perform the navigation on the one-dimensional path, wherein the SLAM method identifies an improvement in a first clinical indicator while minimizing a degradation in a second clinical indicator.

31. The computing system of claim 24, wherein: The set of additional weights in the submanifold space corresponds to a level set of the optimality condition for the radiation therapy problem.

32. The computing system of claim 24, wherein: The derivatives are provided by automatic differentiation.

33. The computing system of claim 24, wherein: At least a portion of the manifold space has a non-differentiable portion.

34. The computing system of claim 33, wherein: The non-differentiable portion of the manifold space is smoothly approximated using a least squares method.

35. The computing system of claim 33, wherein: The barrier formulation is used to convert a constrained problem into an unconstrained problem by including the constraints as terms in the objective function and associating a penalty for violating the constraints.

36. The computing system of claim 24, wherein: The plurality of criteria for the radiation therapy question corresponds to clinical preferences, and wherein at least one criterion of the plurality of criteria relates to a specific anatomical region for receiving the radiation therapy from the radiation therapy machine.

37. The computing system of claim 24, wherein: The one or more processors are further configured to: A representation of a solution space is generated based on the plurality of solutions to the radiation therapy problem, wherein the solution space is a Pareto surface including a set of Pareto optimal solutions.

38. The computing system of claim 24, wherein: The one or more processors are configured to perform the following operations: generating a display of a graphical user interface configured to provide functionality for configuring the treatment plan; and Information associated with the additional solutions is displayed in the graphical user interface.

39. The computing system of claim 24, wherein: The one or more processors are configured to perform the following operations: receiving a selection of the additional solution; Wherein the treatment planning parameters are generated based on the selection of the additional solution.

40. The computing system of claim 39, wherein: The additional solutions provide a warm start to identify a solution for the treatment plan, the one or more processors are configured to perform the following operations: receiving an optimization of the additional solution; Wherein the treatment plan parameters are generated based on the optimization of the additional solution.

41. The computing system of claim 24, wherein: The treatment planning parameters for the radiation therapy include a set of treatment delivery parameters corresponding to the capabilities of the radiation therapy machine.

42. The computing system of claim 41, wherein: The treatment plan is for providing the radiation therapy using a gamma knife, and wherein the set of treatment delivery parameters includes a set of isocenters for delivering the radiation therapy.

43. The computing system of claim 41, wherein: The set of treatment delivery parameters also includes a timing for delivering the radiation therapy, and a collimator sequence for delivering the radiation therapy.

44. The computing system of claim 41, wherein: The treatment plan is used to deliver the radiation therapy using a linear accelerator or a magnetic resonance (MR)-linear accelerator radiation therapy machine.

45. The computing system of claim 44, wherein: The treatment plan is for providing volumetric arc modulated therapy (VMAT) or intensity modulated radiation therapy (IMRT) to the radiation therapy, and wherein the set of treatment delivery parameters includes: a set of arc control points for one or more arcs, a fluence field, a gantry speed, and a dose rate along the one or more arcs.

Images