Using organ similarity metrics to determine optimization strategies in adaptive radiation therapy

Abstract

Systems, devices, and methods for determining an optimization strategy in adaptive radiation therapy using similarity metrics, and systems and methods for an auto-adaptive workflow for using an MLC blade configuration selected based on calculated similarity metric values, and the processing plan of the current processing link is automatically adapted and optimized.

Description

Technical Field

[0001] The present disclosure relates generally to adaptive radiation therapy and, more particularly, to systems, methods, and apparatus for optimizing radiation treatment plans in volumetric modulated arc therapy (VMAT) using an organ similarity metric. Background Art

[0002] Radiation therapy involves medical procedures that use external beam radiation to treat pathological anatomy (tumors, lesions, vascular malformations, neurological conditions, etc.) by delivering a prescribed dose of radiation (X-rays, gamma rays, electrons, protons and / or ions) to the pathological anatomy while minimizing radiation exposure to surrounding tissues and critical anatomical structures.

[0003] In general, a complete radiation therapy planning and treatment workflow includes several phases: a treatment planning phase, a treatment delivery phase, and a monitoring and evaluation phase, in which the progress of the treatment (eg, dose accumulation) is monitored.

[0004] In the treatment planning stage, an accurate three-dimensional (3D) map of the anatomical structure in the region of interest (head, body, etc.) is first constructed using any one of computed tomography (CT), cone beam computed tomography (CBCT), magnetic resonance imaging (MRI), positron emission tomography (PET), 3D rotational angiography (3DRA) or ultrasound techniques (or a combination thereof). This determines the exact coordinates of the target within the anatomical structure, i.e., locates the tumor or abnormality in the body and defines its exact shape and size. This is followed by a prescription step in which the motion path for the radiation beam is calculated to deliver an acceptable dose distribution found by the radiation oncologist taking into account various medical constraints. A team of experts then develops a treatment plan using specific computer software to optimally irradiate the tumor and minimize the dose to surrounding normal tissue by designing radiation beams that converge on the target area from different angles and planes. The treatment plan is then evaluated against the prescription.

[0005] In the treatment delivery phase, the radiation treatment plan is executed. During this phase, the radiation dose is delivered to the patient according to the prescribed treatment plan. Typically, the treatment plan is delivered to the patient through a series of radiation treatments called fractions. However, there are many factors, such as differences in the patient's posture position, changes that may occur if the patient's tumor regresses or if the patient loses weight during the therapy, and uncertainties introduced by movement, which may cause differences between the prescribed radiation dose distribution and the actual dose delivered (i.e., the actual dose delivered to the target during the radiation treatment). These anatomical and physiological changes may cause the target volume and surrounding anatomical structures and organs to move during the therapy and cause size and shape changes. In this way, executing or continuing to execute the original treatment plan may cause the actual received dose distribution to be different from the planned distribution, and therefore cause a dose reduction of the target volume and / or an increase in the dose of the risk organ (OAR). Therefore, during the treatment delivery phase, the treatment plan can be adapted according to the images of the day to better reflect the current situation. This involves modifying the original treatment plan based on the image data subsequently acquired to match the new position and shape of the target volume and surrounding anatomical structures.

[0006] Adaptive radiation therapy is a process by which, using subsequent images, the initial / original treatment plan can be adjusted to offset these anatomical changes. The adaptive radiation therapy process is a closed-loop radiation treatment process in which the initial treatment plan can be modified using systematic measurement feedback. Adaptive radiation therapy improves radiation treatment by systematically monitoring treatment changes and re-optimizing the initial treatment plan by incorporating them during the treatment process.

[0007] Adaptive radiation therapy can be performed on three different time scales, namely offline between treatment fractions, online immediately before a treatment fraction, and real-time during a treatment fraction.

[0008] During offline adaptive therapy, new images (e.g., CT or CTBC images) of the patient are acquired during each treatment fraction, before or after each fraction in the fraction, and these images are evaluated to determine the multi-day position of the target volume. Based on this, a new treatment plan can be developed to better reflect the extent of the motion of the target volume.

[0009] In an online (on-couch) real-time adaptive therapy process, the radiation therapy system can be used to verify or adjust the treatment plan before fractionation for treatment delivery. Thus, the treatment delivery can be modified simultaneously using the imaging system to reflect changes in the patient's anatomy.

[0010] During couch-based adaptive therapy, the radiation therapy system may also be used during treatment fractions. Thus, couch-based adaptive radiation therapy allows the initial treatment plan to be adjusted based on tumor and anatomical changes while the patient is on the treatment table.

[0011] Generating an adapted plan during the treatment delivery phase of an adaptive treatment workflow can be time consuming and cumbersome. Although some of the steps involved in the plan adaptation process have been automated to assist and reduce the workload of clinical users, it remains difficult to properly automate the optimization of treatment plans, especially in the fields of intensity modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT).

[0012] Volumetric modulated arc therapy (VMAT) refers to the technique of delivering intensity modulated fields by rotational delivery using a conventional linear accelerator equipped with a binary multileaf collimator (MLC). Compared to conformal arc, the treatment field does not necessarily conform to the target volume at every angle. Instead, the effective intensity modulated field is delivered by arc sectors. Therefore, VMAT treatment planning aims to determine the optimal trajectory of MLC leaves, dose rate and gantry angle.

[0013] Optimization is an iterative process in which the user attempts to specify planning objectives in the form of dose or biological objectives to create an ideal dose to target structures / target volumes and minimize doses to critical structures. Optimization algorithms typically require a starting point (input metric) that specifies the preferred kind of dose distribution. Typically, the starting point is a set or template of clinical goals (CGs) that specify the preferred dose distribution (e.g., how much dose is needed to the target structure / target volume and what is the upper limit dose in the risk organ), or a knowledge-based model that guides the optimization process (e.g., the model estimates the achievable dose distribution for each structure).

[0014] However, current volumetric modulated arc therapy (VMAT) plan optimization algorithms are often unstable to small changes in input data. As a result, the daily treatment plan may differ entirely from the initial treatment plan. This impact is even greater in cases where the optimization process is based on meeting prioritized clinical goals.

[0015] Attempts have been made to solve this problem by using the MLC blade positions from a preliminary plan determined for a planning image (taken before the actual daily treatment session images). However, this approach has disadvantages. If the deformation between the planning image and the daily treatment session image is too large, the MLC blade sequence obtained in the preliminary plan may already be in the basin of some unacceptable local minimum, in which case the VMAT optimization algorithm will not be able to find a better solution (i.e. a better treatment plan) because the scanning of the solution space is limited.

[0016] Another attempt to address this problem is to deform the initially planned MLC leaf sequence based on the deformation field between the planning image and the daily treatment session images. However, performing such a deformation is not well defined, as there is no guarantee that deformation of the MLC leaf positions will result in a good treatment plan. Some target coverage may be improved, but critical organ doses may not be well controlled.

[0017] To speed up the optimization process in couch-based adaptive therapy, the treatment adaptation functionality currently used completes the optimization by starting from standard / generic initial conditions, rather than taking a previously generated treatment plan as a starting point.

[0018] Therefore, a system and method are needed to automatically optimize the adapted treatment plan in an adaptive workflow to address these deficiencies in the optimization process. Summary of the invention

[0019] Systems and methods configured to use a similarity metric to determine an optimization strategy in adaptive radiation therapy are described herein.

[0020] Further described herein are systems and methods that enable an automated process for adaptive radiation therapy sessions to generate and optimize treatment plans using an optimization algorithm that includes optimization parameters generated based on control points that are automatically selected based on a similarity metric.

[0021] Further disclosed are systems and methods that enable: obtaining an instruction set including a plan quality value for a preliminary treatment plan for a planning patient model; generating a treatment session patient model using information from the instruction set; determining a similarity measure between a target structure of the planning patient model and a corresponding target structure of the treatment session patient model; and generating an adapted treatment plan for the treatment session patient model by: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan; the treatment plan generation algorithm including optimization parameters generated based on control points, the control points being automatically selected based on the determined similarity measure; and generating an adapted treatment plan by optimizing the current plan based on the plan quality value of the preliminary treatment plan.

[0022] Further disclosed are the following systems and methods, which implement a method for adapting a preliminary radiation treatment plan, the preliminary radiation treatment plan being used for a current treatment phase of an adaptive radiation therapy phase, the method comprising: obtaining an instruction set, the instruction set comprising a plan quality value of a preliminary treatment plan for a patient; using the instruction set, executing a series of automated steps to generate a current treatment phase patient model; determining a similarity measure between a target structure of the current treatment phase patient model and corresponding target structures of a plurality of previously obtained treatment phase patient models; determining a previously obtained treatment phase patient model having a highest similarity measure value; and generating a current treatment phase patient model by the following operations An adapted treatment plan for a stage patient model: generating a current plan using a preliminary treatment plan as input to a treatment plan generation algorithm; the treatment plan generation algorithm comprising optimization parameters generated based on control points, the control points being automatically selected based on a determined similarity metric; and generating an adapted treatment plan by optimizing the current plan based on a plan quality value of the preliminary treatment plan, wherein the control points comprise an MLC blade configuration, and wherein when a highest similarity metric value is below a predetermined threshold, a standard MLC blade configuration is selected, and when a highest similarity metric value is above a predetermined threshold, an MLC blade configuration of a treatment plan of a previous treatment stage patient model having the highest similarity metric value is selected.

[0023] Also disclosed are systems and methods for implementing an automated workflow for an adaptive radiotherapy session for a patient, comprising: a computer processing system configured to: obtain an instruction set including a plan quality value for a preliminary treatment plan for the patient; and use the instruction set to perform a series of automated steps to: generate a current treatment session patient model; and generate an adapted treatment plan for the treatment session patient model by: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan, optimizing the current plan based on optimization parameters, the optimization parameters being generated based on control points, the control points being selected based on a similarity metric determined between a target structure of the current treatment session patient model and a corresponding target structure of a planning patient model or one or more previously obtained treatment session patient models, wherein the control points include an MLC leaf configuration, and wherein when the highest similarity metric value is below a predetermined threshold, a standard MLC leaf configuration is selected, and when the highest similarity metric value is above a predetermined threshold, the MLC leaf configuration of the treatment plan of the previous treatment session patient model having the highest similarity metric value is selected.

[0024] Also disclosed is a system comprising a computer processing device configured to execute a sequence of programmed instructions embodied on a computer-readable storage medium, the execution of which causes the system to perform the method steps disclosed herein.

[0025] Also disclosed is a non-transitory computer-readable storage medium having embodied thereon a sequence of programmed instructions for generating daily treatment images to be used in adaptive radiation therapy, and a computer processing system that executes the sequence of programmed instructions embodied on the computer-readable storage medium. Execution of the sequence of programmed instructions may cause the computer processing system to perform the automatic treatment planning and optimization process described herein.

[0026] Objects and advantages of embodiments of the disclosed subject matter will become apparent from the following description when considered in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1A-1B is a simplified schematic diagram of a radiation therapy system in accordance with various embodiments of the disclosed subject matter.

[0028] Figure 2 is a simplified illustration of using the radiation therapy system of FIG. 1 for couch-based adaptive radiation therapy, in accordance with various embodiments of the disclosed subject matter.

[0029] Figure 3 is a workflow diagram for couch-based adaptive radiation therapy according to various embodiments of the disclosed subject matter.

[0030] Figure 4-Figure 5 is a flow chart of a process for generating a planning patient model in accordance with various embodiments of the disclosed subject matter.

[0031] Figure 6 is a flow chart of a process for generating a preliminary treatment plan in accordance with various embodiments of the disclosed subject matter.

[0032] Figure 7-10 is a flow chart of a process for generating a treatment session patient model in accordance with various embodiments of the disclosed subject matter.

[0033] Fig.11 is a flow chart of a process for determining a similarity measure in accordance with various embodiments of the disclosed subject matter.

[0034] Fig.12 is a flow chart of a process for rigid registration between images in accordance with various embodiments of the disclosed subject matter.

[0035] Fig.13 is a flow chart of a process for determining a treatment isocenter in accordance with various embodiments of the disclosed subject matter.

[0036] Figure 14-15 is a diagram of different similarity coefficient calculation processes according to various embodiments of the disclosed subject matter.

[0037] Fig.16is a process flow diagram of a process for automatic generation and optimization of a fitting plan based on a similarity metric in accordance with various embodiments of the disclosed subject matter.

[0038] Fig.17 is a flow chart of a process for generating a scheduled treatment plan in accordance with various embodiments of the disclosed subject matter.

[0039] Fig.18 is a process flow diagram of a process for automatic generation and optimization of a fitting plan based on a similarity metric in accordance with various embodiments of the disclosed subject matter.

[0040] Fig.19 is a process flow diagram for a similarity coefficient calculation process according to various embodiments of the disclosed subject matter.

[0041] Fig. 20 is a process flow diagram of a process for automatic generation and optimization of a fitting plan based on a similarity metric in accordance with various embodiments of the disclosed subject matter. DETAILED DESCRIPTION

[0042] refer to Figure 1A and Figure 1B , shows an exemplary radiation therapy system 100, which can be used in Figure 2 , and can deliver radiation according to a treatment plan determined using the techniques described herein. The radiation therapy system 100 can provide radiation to a patient 110 positioned on a treatment couch 112, and can allow for implementation of various radiation dose verification protocols. Radiation therapy can include photon-based radiation therapy, particle therapy, electron beam therapy, or any other type of treatment therapy.

[0043] In one embodiment, the radiation therapy system 100 may include a radiation treatment device 101, such as but not limited to a LINAC, which is operable to generate one or more beams of megavoltage (MV) X-ray radiation for treatment. The LINAC may also be operable to generate one or more beams of kilovoltage (kV) X-ray radiation, for example, for patient imaging. The system 100 has a gantry 102 supporting a radiation treatment head 114 having one or more radiation sources 106 and various beam modulation elements, such as but not limited to a flattening filter 104 and a collimation assembly 108. The collimation assembly 108 may include, for example, a multi-leaf collimator (MLC), an upper jaw and a lower jaw, and / or other collimation elements. The collimation assembly 108 and / or the flattening filter 104 may be positioned within the radiation beam path by a corresponding actuator (not shown) that may be controlled by the controller 200.

[0044] The gantry 102 can be an annular gantry (i.e., it extends through a full 360° arc to create a complete ring or circle), but other types of mounting arrangements can also be used. For example, a static beam or a C-shaped partial annular gantry or robotic arm can be used. Any other framework capable of positioning the treatment head 114 at various rotational and / or axial positions relative to the patient 110 can also be used.

[0045] In one embodiment, the radiotherapy device is an MV energy intensity modulated radiotherapy (IMRT) device. The intensity distribution in such a system is customized for the treatment requirements of individual patients. The IMRT field is delivered with MLC 108, which can be a computer-controlled mechanical beam shaping device attached to head 114 and includes an assembly of metal fingers or leaves. For each beam direction, an optimized intensity distribution is achieved by sequential delivery of various subfields with optimized shapes and weights. From one subfield to the next, the leaves can move as the radiation beam is turned on (i.e., dynamic multi-leaf collimation (DMLC)) or as the radiation beam is turned off (i.e., segmented multi-leaf collimation (SMLC)).

[0046] Alternatively or additionally, the radiotherapy device 101 may be a tomotherapy device, a helical tomotherapy device, or a simplified intensity modulated arc therapy (SIMAT) device, a volumetric modulated arc therapy (VMAT) device, or a volumetric high definition (or super arc) therapy (HDRT) device. In fact, any type of IMRT device may be used as the radiotherapy device 101 of the system 100, and may also include on-board volume imaging, which may be used to generate in-process image data generated during the processing session.

[0047] For example, embodiments of the disclosed subject matter may be applied to an image-guided radiation therapy (IGRT) device that uses cross-sectional images of a patient's internal anatomy taken during a radiation therapy treatment session (i.e., in-process images) to provide information about the patient's position. During radiation treatment, frequent two-dimensional or three-dimensional imaging is used to guide the therapeutic radiation using imaging coordinates of the actual radiation treatment plan. This ensures that the patient's position in the radiation treatment system is the same as planned and that the patient is properly aligned during treatment. Although the IGRT process involves conformal radiation treatment guided by specialized imaging tests performed during the planning phase, it does rely on imaging patterns from the planning process as reference coordinates for positioning the patient 110 during treatment. Therefore, associated with each image-guided radiation therapy system is an imaging system for providing in-process (process session) images for setting up the radiation delivery program.

[0048] The in-process images may include one or more two-dimensional or three-dimensional images (typically x-rays) acquired at one or more different points during the treatment. There are multiple methods for acquiring in-process images. In some methods, different independent imaging systems and / or imaging methods are used to acquire the pre-process and in-process images. For example, 3D IGRT may include positioning a cone beam computed tomography (CBCT) data set with a planning computed tomography (CT) data set, and 2DIGRT may include matching a planar kilovoltage (kV) radiograph or megavoltage (MV) image with a digitally reconstructed radiograph (DRR) obtained from the planning CT.

[0049] Each type of radiation therapy device may be accompanied by a corresponding radiation plan and radiation delivery procedure.

[0050] The controller 200 (which may be, but is not limited to, a graphics processing unit (GPU)) may include a computer having appropriate hardware such as a processor, and an operating system for running various software programs and / or communication applications. The controller 200 may include a software program that operates to communicate with the radiation therapy device 101, which is operable to receive data from external software programs and hardware. The computer may also include any appropriate input / output (I / O) device 210, which may be suitable for allowing communication between the controller 200 and a user (e.g., medical staff) of the radiation therapy system 100. For example, the controller 200 may be provided with an I / O interface, a console, a storage device, a memory, a keyboard, a mouse, a monitor, a printer, a scanner, and a department information system (DIS) such as a communication and management interface (DICOM) for storing and transmitting medical imaging information and related data, and enabling integration of medical imaging devices such as scanners, servers, workstations, printers, network hardware, etc.

[0051] Alternatively or additionally, the I / O device 210 may provide access to a network (not shown) for transmitting data between the controller 200 and remote systems. For example, the controller 200 may be networked with other computers and radiation therapy systems via the I / O 210. The radiation therapy system 100, the radiation treatment device 101, and the controller 200 may communicate with the network as well as databases and servers, such as a dose calculation server (e.g., a distributed dose calculation framework), a similarity determination system 300, and a treatment planning system 400. The controller 200 may also be configured to transmit medical image-related data between different medical equipment.

[0052] The system 100 may also include multiple modules (e.g., as part of the controller 200, or as separate modules within the system 100, or integrated into other components of the system 100) that contain programmed instructions that, when executed, cause the system 100 to perform various functions related to adaptive radiation therapy or other radiation treatments, as discussed herein. For example, the system 100 may include: a treatment planning module, which is operable to generate a treatment plan for the patient 110 based on multiple data input into the system by medical personnel; a patient positioning module, which is operable to position and align the patient 110 relative to a desired position (such as the isocenter of the gantry) for a specific radiotherapy treatment; an image acquisition module, which is operable to instruct the radiotherapy system and / or imaging device to acquire images of the patient 110 before the radiotherapy treatment (i.e., pre-treatment / benchmark / planning images used for treatment planning and patient positioning) and / or acquire images of the patient 110 during the radiotherapy treatment (i.e., link images in treatment / daily treatment), and to instruct the radiotherapy system 100 and / or imaging device 101 or other imaging devices or systems to acquire images of the patient 110.

[0053] The system 100 may also include: a radiation dose prediction module, which is operable to predict the dose to be delivered to the patient 110 before the start of the radiation treatment therapy; a dose calculation module, which is operable to calculate the actual dose delivered to the patient 110 during the radiation therapy treatment; a treatment delivery module, which is operable to instruct the radiation therapy device 100 to deliver the treatment plan to the patient 110; an association module, which is operable to correlate the planning images with the images of the in-process / daily treatment sessions obtained during the radiation therapy; a calculation module, which is operable to reconstruct a three-dimensional target volume based on the in-process images; an analysis module, which is operable to calculate displacement measurements; and a feedback module, which is operable to instruct the controller to stop the radiation therapy in real time based on a comparison of the calculated displacement with a predetermined threshold (range).

[0054] The system 100 may also include: one or more contour generation modules, which are operable to generate contours of target volumes and other structures in pre-processing (planning, reference) images and in-processing (processing session) images; an image registration module, which is operable to register a pre-processing image with a subsequent in-processing image; a dose calculation module, which is operable to calculate a cumulative dose; a contour propagation module, which is operable to propagate contours from one image to another; a contour verification module, which is operable to verify the generated contours; a registration deformation vector field generation module, which is operable to determine a deformation vector field (DVF) as a result of the image deformation process; a similarity metric determination / calculation module, which is operable to determine / calculate a similarity metric between the target structure of the planning image and the target structure of the in-processing / daily processing session image, or between the target structure of the planning image and each daily processing session image in the daily processing session images; and a comparison module, which is operable to compare the similarity metric value with a predetermined threshold. The system 100 may also include modules for electron density map generation, equivalent dose distribution generation, dose volume histogram (DVH) generation, image synchronization, image display, treatment plan generation, treatment plan optimization, automatic optimization parameter generation, update and selection, and adaptive instruction and processing information transmission. For example, the modules can be written in C or C++ programming languages. Computer program code for implementing operations as described herein can be written in any programming language, such as C or C++ programming languages.

[0055] Figure 3 An automated workflow S100 implemented via module 220 of system 100 is illustrated, which includes an organ similarity metric as a basis for on-couch adaptive radiation treatment planning ( Figure 2 The input of the automatic optimization process in ).

[0056] The automated adaptive workflow S100 initiates the generation of a treatment plan for the system 100 via the treatment planning system 400. The treatment planning system 400 may be an independent stand-alone treatment planning system, or may be configured as a part of the system 100 as a treatment planning module. The treatment planning system 400 may be used to generate one or more treatment plans (a preliminary / initial treatment plan, and one or more treatment plans adapted to current / daily conditions) based on image data (such as CT or CBCT image data previously captured and stored in a storage device, or CT or CBCT image data generated by the system 100 before or during a treatment session).

[0057] The treatment planning system 400 includes a patient model generation module 400A for generating a planning patient model 410 / 410 ′, and includes a treatment plan generation module 400B for generating an initial / preliminary treatment plan 420 for the planning patient model 410 / 410 ′.

[0058] In the usual planning process, qualified medical personnel (physicians) manually draw contours on one or more of the initial planning images (such as CT images 310). These contours depict the malignant tumor to be irradiated, as well as one or more other structures, such as organs, tissues, etc. that are susceptible to substantial damage from radiation exposure. The planning image can also be semi-automatically or automatically segmented to depict the malignant tumor to be the irradiation target and any surrounding critical structures (OARs) whose irradiation should be limited. The usual depiction for malignant tumors includes the total target volume (GTV), the clinical target volume (CTV), and the planning target volume (PTV). The GTV determines the anatomical region with the highest tumor cell density and the highest prescription dose required. The GTV is the location and range of the total tumor, that is, the location and range that can be seen, palpated, or imaged. The CTV contains the GTV plus the boundaries of the subclinical disease spread, so the CTV cannot be fully imaged. The CTV is a volume that must be fully treated to achieve a cure. The PTV allows for uncertainty in planning or treatment delivery. This is a geometric concept designed to ensure that the radiation therapy dose is actually delivered to the CTV. Therefore, the PTV is used to compensate for treatment setup uncertainties by volumetric expansion of the CTV borders.The planning images may also illustrate soft tissues, object structures, organs, vessels, bones, etc.

[0059] One or more anatomical structures of interest (e.g., body contours) and one or more influencer structures (contours of organs and / or non-volumetric structures that influence one or more of the shape, size, and position of one or more target structures in the target CT image 310) may also be generated on the planning CT image 310.

[0060] Influencers are structures that influence one or more of the shape, size, and position of one or more of the primary targets, nodal targets, major organs, and other anatomical structures of interest during online fitting. These influencers can include structures that typically move from day to day and / or exhibit large deformations and / or motions (such as organs), as well as other non-volumetric structures (such as points or 2D lines that describe the anatomy).

[0061] like Figure 4 As shown in FIG. 4 , the contour drawing module 402A draws / depicts the contours of the target (eg, organs to be irradiated and organs to be spared) on the planning CT image 310 to obtain the planning CT image 310 including the target structure 409. Figure 5As shown in , the planning patient module 410' may further include an influencer structure 413A, which is obtained by the contour drawing module 402A drawing not only the contours of the target 411A' to be irradiated and the target 412A' to be avoided but also the contours of one or more influencers.

[0062] The target structures and influencer structures depicted on the planning CT image 310 can be evaluated by qualified medical personnel (physician) using the contour evaluation module 403A on the computer display 401 A. If accepted, the planning CT image 310 including the target structures / influencer structures / anatomical structures of interest becomes the patient model 410 / 410'.

[0063] The planning patient model 410 / 410', the planning image (i.e., the CT image 310), the target structures on the planning image (i.e., the planning target structures 411 / 411', 412 / 412'), and information about the planning CT image 310 and the shape and position of the planning target structures in the planning CT image 310, information about the influencer 413, and any other depicted anatomical structures are sent to the system 100 via I / O 200, DICOM, or an imaging console, or are made available to the system 100 via I / O 200, DICOM, or an imaging console.

[0064] In order to generate a treatment plan, a medical staff (physician) generates a list of treatment parameters 230', such as but not limited to the target to which radiation is to be maximized, the target to which radiation is to be minimized, and other parameters and instructions related to the specific system 100 and imaging device 101 that will deliver the treatment plan to the patient 110. The physician also specifies a preferred dose distribution for the target structure. The dose distribution is represented as a set or template of clinical goals (CG), which are appropriate goals for the radiation dose of the patient's treatment. These clinical goals (CG) can be given, for example, in the form of an average radiation dose (in Gray) to the target structure and a dose that is prohibited to be exceeded for a certain volume of organs (such as risk organs (OARs)). However, clinical goals can also be given in other dimensions, which are not in the form of radiation doses to target structures and doses to organ volumes. Each goal in each given goal can also be prioritized, and the priority describes the importance of meeting one goal compared to another goal. This set is called a prioritized set of clinical goals (prioritized CG). Each clinical goal can be expressed as a quality metric Q and its associated goal value. An exemplary prioritized set of clinical objectives is:

[0065] Objective 1: The target (PTV) must receive 50 Gy: Priority 1

[0066] Objective 2: Organ at risk X (OARx) must receive less than 25 Gy: Priority 2

[0067] Objective 3: Organ at risk Y (OARy) must receive less than 30 Gy mean dose: Priority 3

[0068] The treatment plan generation module 400B then transforms the cost function (C ) of the dose distribution that defines the prioritized set of clinical goals (CG) into a single cost function (C G ) is minimized to automatically generate an optimized plan (i.e., preliminary plan 420) for the planned patient module 410 / 410'. There are many algorithms that can be applied to minimize the cost function, including but not limited to calculating the gradient of the cost function. The solution of the optimization process results in the determination of treatment parameters for the treatment plan candidate 420'.

[0069] For VMAT planning, at the beginning of the treatment planning process, a plurality of control points (CPs) are also specified for the beam trajectory, which take into account the beam shaping elements of the system 100. Each control point (CP) is associated with a set of treatment parameters, including but not limited to a set of MLC leaf positions, MLC shape, gantry rotation speed, gantry position, dose rate, cumulative number of monitoring units (MUs) delivered to the control point, and / or any other parameters such as maximum MLC leaf speed, minimum leaf gap or stagger constraints for moving MLC leaves, maximum leaf travel per degree of gantry rotation, etc.

[0070] The number and location of the control points (CP) may be set in any convenient manner, such as but not limited to by using the treatment planning software or by a system operator. In an exemplary embodiment, the beam trajectory may include a single 180 degree arc trajectory and approximately 177 sequential control points (CP), which means that there are 177 configurations that the LINAC should conform to in order to deliver the planned treatment. Based on the treatment parameters, the dose distribution within the treatment volume may be calculated for each control point (CP) by any number of techniques, such as but not limited to pencil beam convolution or any other appropriate algorithm, and the radiation dose distribution for each (CP) may be associated with the corresponding gantry angle, MLC configuration, and monitor unit (MU). Thus, during treatment delivery, the extracted (CP) parameters may be associated with the corresponding calculated dose distribution for each control point (CP). Thus, the treatment plan candidate 420' for the VMAT plan determines the optimal trajectory, dose rate, and gantry angle of the MLC blade that meets the prioritized set of clinical objectives.

[0071] The physician may also evaluate the treatment plan candidates 420' at 403B. The accepted treatment plan candidates become the preliminary treatment plans 420. The preliminary treatment plans 420 and associated parameters (i.e., initial MLC leaf configuration 430, clinical objectives, dose specifications, clinical objective values, treatment plan 3D doses, how the plan is optimized, control points, etc.) are stored in a storage device of the treatment planning system 400 and later retrieved by the system 100 to operate the system 100 to deliver radiation treatment according to the selected radiation treatment plan.

[0072] The physician also formulates a set of adaptive instructions 230 , which is a list of parameters / instructions / information describing the adaptive treatment intent, ie, a 4D description of the planned treatment for the patient 110 . The set of adaptive instructions may include: information about the planned dose specifications (i.e., Rx prescription), whether to use adaptive IGRT therapy or standard IGRT therapy, specified clinical objectives (such as but not limited to target dose coverage and (OAR) risk dose limits, planned clinical objective values), planning images 310, supporting images and their corresponding registration information (PET, MRI, etc.), planning patient model 410 / 410' (i.e., contours of planning structures, such as target volumes, OARs and other internal dose-derived structures on the planning images), preliminary treatment plan 420, initial MLC leaf configuration 430 associated with the preliminary treatment plan 420, preliminary treatment plan 3D dose (i.e., RT 3D dose), a list of target structures of the planning CT image 310 (target volumes, OARs, influencer structures, body contours), a list of influencer structures for different treatment sites, information about the shape and position of the planning structures on the planning images, and any information about how the preliminary treatment plan is optimized.

[0073] The goal of adaptive radiation therapy is to appropriately modify the radiation treatment plan (i.e., preliminary treatment plan 420) to account for temporal changes in the patient's anatomy. As such, images (such as daily CBCT images) acquired (500) at the treatment site during a treatment session (i.e., treatment session images 510) are sent to the treatment planning system 400, where the preliminary treatment plan 420 can be adapted to the current anatomy via deformable registration software and sent back to the system 100 for delivery.

[0074] To adapt the preliminary treatment plan 420, an accurate treatment-session patient model 610 (i.e., the outline of the target on the current anatomy) first needs to be generated, and then the preliminary treatment plan 420 is appropriately modified to accommodate the new anatomy and quickly evaluated for application to the patient 110. The on-couch adaptive workflow 220 generates the treatment-session patient model 610 in a step-by-step manner through a series of automated steps guided by a set of adaptive instructions 230. The set of instructions 230 is also used to guide the generation (800), selection, and delivery (900) of the treatment plan that best fits the patient's current anatomy.

[0075] like Figure 2 and Figure 7 As shown in FIG, a radiologic technician (RTT / MTA) who is experienced and well-trained in reviewing anatomy and plan selection and is responsible for delivering adaptive treatment to a patient (i.e., user / fitter) performs a first level treatment modification by placing the patient (211) on a treatment couch and moving the patient to an imaging position (see FIG. Figure 2 ).

[0076] After patient setup (211), one or more treatment session images (510) of a portion of interest of the patient 110 are captured (500) using the radiation imaging device 101. In an exemplary embodiment, the treatment session images are 3D or 4D CBCT scans, such as obtained during a treatment session by irradiating a region of interest of the patient 110 with radiation 120. The treatment session images 510 may display the bone structure of the patient 110, but may not include any depiction of a target volume or other structures.

[0077] The RTT / MTA then initializes (600) the generation of contours of the planning structures (411A, 412A) on the processing link image 510 via the contour drawing module 403A to obtain a processing link patient model 610 (i.e., a processing link CBCT image 510 including the target structures of the planning patient link 410). This can be accomplished by propagating (601) the planning target structures (411A, 412A) from the planning patient model 410 onto the processing link image 510 to obtain (602) a processing link CBCT image 510 including the propagated target structures (411B, 412B). The RTT / MTA can evaluate the processing link image thus obtained, and if accepted, the image and associated image data will become the processing link patient model 610, as shown in FIG. Figure 7-Figure 8 As shown in .

[0078] Alternatively, the RTT / MTA initializes (603) the delineation of the influencers included in the set of adaptive instructions 230 on the processing link image 510 to obtain a processing link CBCT image 510 (604) including the influencer structures, and then the RTT / MTA initiates (605) the propagation of the planning target structures (411A', 412A') included in the planning link model 410A' to obtain (606) the processing link CBCT image 510 containing the propagated target structures (411B', 412B'). That is, the RTT / MTA initiates the propagation of the target volume (CTV, PTV, etc.), OARs, and other anatomical structures of interest in addition to the influencer structure 413A present in the planning link model 410A'. The RTT / MTA can evaluate the obtained processing link image (606), and if accepted, the image and associated image data become the processing link patient model 610', as shown. Fig. 9 and Fig.10 As shown in .

[0079] The propagation of the target structures (411A, 412A, 411A', 412B') from the planning CT image 310 of the planning patient model 410 / 410' can be accomplished using a structure-guided deformable registration algorithm (Structure-Guided DIR), which registers the image data of the planning patient model 410A / 410A' with the image data of the processing-stage patient model 610 / 610', and thereby generates one or more deformable vector fields (DVFs), which are used to propagate the target volume (CTV, PTV, etc.), OARs and other anatomical structures of interest from the planning CT image 310 of the planning patient model 410A / 410A' to the processing-stage CBCT image 510 of the processing-stage patient model 610 / 610' using the contour propagation module 403A.

[0080] When using influencers, the guided deformable registration is a deformable registration that is guided by the relationship between the influencer structure 413A present in the planning CT image 310 of the planning patient model 410A′ and the influencer structure 413B generated in the processing stage CBCT image 510 .

[0081] After obtaining the planning patient model 410 / 410' (including the target structures (411A / 411A', 412A / 412A') on the CT image 310) and the processing-stage patient model 610 (including the target structures (411B / 411B', 412B / 412B') on the processing-stage CBCT image 510), process S100 proceeds to determining (700) a similarity measure 710 between the planning patient model 410 / 410' and the processing-stage patient model 610 / 610'.

[0082] Determination of the similarity metric includes calculating a coefficient / index / metric measuring the similarity between the planned target structure (411A / 411A', 412A / 412A') and the corresponding processing link target structure (411B / 411B', 412B / 412B').

[0083] An exemplary similarity metric is the Sorensen-Dice score (DSC i ), the score can be calculated using the following equation:

[0084]

[0085] in and is the set of points belonging to the target structure i in the first image and in the second image, for which planning has been completed for the first image (i.e., the target structures 411A, 411A' / 412A, 412A' in the planning image of the planning patient model 410 / 410', or the target structures from the previous processing link), and for which planning has not yet been completed for the second image (i.e., the same target structures on the processing link image propagated to the processing link patient model 411B / 411B', 412B / 412B').

[0086] Another exemplary similarity metric is the Jaccard index J(A, B), which can be calculated using the following equation:

[0087]

[0088] Wherein A is the set of points belonging to the target structure in the first image, for which planning has been completed (i.e., the target structures 411A, 411A' / 412A, 412A' in the planning image of the planning patient model 410 / 410', or the target structures from the previous processing link), and B is the set of points belonging to the same target structure in the second image, for which planning has not yet been completed (i.e., the same target structure on the processing link image propagated to the processing link patient model 411B / 411B', 412B / 412B').

[0089] Another exemplary similarity metric is the Hausdorf metric, which measures how far two subsets of the metric space are from each other. In this example, it measures how far the set of points belonging to the target structure in the first image that has been planned (i.e., the target structure 411A, 411A' / 412A, 412A' in the planning image of the planning patient model 410 / 410', or the target structure from the previous processing link) is from the set of points belonging to the same target structure in the second image that has not yet been planned (i.e., the same target structure on the processing link image propagated to the processing link patient model 411B / 411B', 412B / 412B').

[0090] Fig.14 An exemplary similarity metric determination 700 process is shown in FIG, wherein a target structure 411A from a planning CT image 310 of a planning patient model 410 is compared with a corresponding target structure 411B from a processing-stage CBCT image 510 of a processing-stage patient model 610. This may be accomplished by overlaying the two corresponding target structures.

[0091] If the contours 411A and 411B do not overlap at all (i.e., zero overlap), the point set of the target structure 411A of the planning image (first image) does not intersect with the point set of the target structure 411B of the processing link image (second image), and the similarity coefficient (SC) is zero (0). If the contours 411A and 411B completely overlap (i.e., completely overlap), the point set of the target structure 411A of the planning image (first image) intersects with all points in the point set of the target structure 411B of the processing link image (second image), and the similarity coefficient (SC) is one (1). If the contours 411A and 411B partially overlap (i.e., there is overlap), some points in the point set of the target structure 411A of the planning image (first image) intersect with some points in the point set of the target structure 411B of the processing link image (second image), and the similarity coefficient (SC) is greater than 0 but less than 1.

[0092] Therefore, the amount of similarity depends on the amount of overlap between the target structure in one image (ie, the planning image) and the same target structure in another image (ie, the processing link image).

[0093] When multiple target structures exist, each target structure in one image is compared with the corresponding target structure in the other image, and a total similarity coefficient is calculated based on the individually calculated similarity coefficients. Fig.15An exemplary similarity metric determination process 700 is illustrated, wherein the planning CT image 310 of the planning patient model 410 includes a first target structure 411A and a second target structure 412A, and the processing link CBCT image 510 of the processing link patient model 610 includes corresponding target structures 411B and 412B. The target structures 411B and 412B are propagated target structures 411A, 412A from the planning CT image 310 to the intraday CBCT processing link image 510.

[0094] like Fig.15 As shown in , the target structure 411A of the planning CT image from the planning patient model 410 is compared with the corresponding target structure 411B of the processing-stage CBCT image from the processing-stage patient model 610, and the target structure 412A of the planning CT image from the planning patient model 410 is compared with the corresponding target structure 412B of the processing-stage CBCT image from the processing-stage patient model 610.

[0095] The overlap amount between the contours 411A and 411B represents a similarity coefficient SC1 between the first target structure in the first image and the second image, and the overlap amount between the contours 412A and 412B represents a similarity coefficient SC2 between the second target structure in the first image and the second image.

[0096] The similarity coefficient (SC) 710 for the scene is calculated as the sum of the similarity coefficients SC1 and SC2, or as the minimum value of the two similarity coefficients SC1, SC2.

[0097] If there are multiple target structures (TS n ), that is, multiple target structures in the planning image (or the previously captured processing link image) of the planned patient model 410 and multiple corresponding propagation target structures on the current processing link image, then the similarity coefficient SC is calculated for each corresponding target structure pair n , and the total similarity coefficient (SC) 710 is the sum of all similarity coefficients (SC n ), or the minimum of all similarity coefficients (701):

[0098] or

[0099] SC=Min(SC1,SC2,...,SC n ).

[0100] The similarity coefficient (SC) can also be calculated (701) by combining similarity coefficients obtained using different similarity metrics. For example, the Sorensen-Dice score DSC and the Jaccard index J(A, B) can be calculated for the target structures 411A and 411B, and the similarity coefficient (SC) can be determined as, for example, the sum of DSC and J(A, B), or Min(DSC, J(A, B)). Obviously, any combination of different similarity metrics can be used to calculate the similarity coefficient (SC) 710.

[0101] For each calculation method of the similarity coefficient (SC) 710, a corresponding threshold value (c) needs to be set in advance. For example, for the similarity coefficient (SC) 710 calculated for a single target structure 411A using the Sorensen-Dice score, the threshold value may be, for example, 0.6, while for multiple targets, if the SC is calculated using the sum of the SCs, the threshold value may be 1.3, or if the SC is calculated as the minimum value in the SCs, the threshold value may be 0.6.

[0102] like Fig.11 As shown in , before determining the similarity measure, a rigid registration algorithm 240 can be applied between the two images, that is, a rigid registration algorithm 240 is applied between the planning CT image of the planning patient model 410 containing the planning target structure (e.g., 411A, 412A) and the processing link CBCT image of the processing link patient model 610 containing the corresponding processing link target structure (e.g., 412A, 412B), so as to correctly align the planning target structure and the processing link target structure before comparison (241).

[0103] The rigid registration algorithm 240 can align the two image sets based on the difference S243 between the processed isocenter position and the planned isocenter position S242, such as Figure 12-13 As shown in Fig.13 As shown in , when the adaptive workflow 220 is initiated, the patient 110 is positioned on the treatment couch 112 in the same manner as a standard radiation therapy treatment. Thus, after positioning the patient 110 on the treatment couch 112, the patient 110 is brought to the isocenter of the system 100, using conventional skin markings to indicate the location of the isocenter.

[0104] When acquiring a processing link image (i.e., for example, a CBCT image) 510, the center of the processing link image 510 corresponds to the acquisition isocenter. If the patient 110 is processed by aligning the acquisition isocenter with the system isocenter, the patient 110 will not be processed correctly because Fig.13 As shown in , the treatment session target volume is not aligned with the planned target volume.

[0105] In order to determine the correct treatment isocenter, after the RTT / MTA accepts the propagated target structure on the treatment link image 510, the system 100 is prompted at S241 to automatically align the planning target structure / volume of the planning patient model 410 / 410' with the treatment link target structure / volume of the treatment link patient model 610 / 610', as shown in FIG. Fig.12 as shown in .

[0106] like Fig.13 As shown in , this alignment provides the difference between the acquisition isocenter and the reference treatment isocenter. Based on the difference, a translation (X, Y, Z) value is calculated, i.e., how much the patient 110 needs to move in the X, Y, and Z directions to align the two isocenters. The calculated translation value is then applied to the acquisition isocenter to determine the treatment isocenter position at S242.

[0107] The treatment isocenter position may also be provided to a plan generation algorithm 402B, which, using information about the preliminary treatment plan 420, may generate a predetermined plan 420', such as Fig.17 The scheduled plan 420' contains information about the new treatment couch 112 location, which will be transmitted to the system 100 if the scheduled plan is selected for treatment. When the scheduled treatment plan contains a new treatment couch location, it forces the RTT / MTA to apply (move the treatment couch 112) the new location.

[0108] After applying the similarity metric and calculating the similarity coefficient (SC) 710 in 700, the similarity coefficient (SC) 710 is compared with the corresponding predetermined threshold value (c) (711). If it is determined in 711 that the calculated similarity coefficient (SC) 710 is greater than the corresponding predetermined threshold value (c), the initial MLC leaf configuration 430 corresponding to the preliminary treatment plan 420 associated with the planned patient model 410 is used as an input metric for the treatment plan adaptation and optimization process 800 to generate an adapted plan 810. If it is determined in 711 that the calculated similarity coefficient (SC) 710 is less than the corresponding predetermined threshold value (c), the standard MLC leaf configuration 440 is used as an input metric for the treatment plan adaptation and optimization process 800 to generate an adapted plan 810. The standard MLC leaf configuration 440 is a predetermined MLC leaf configuration driven by the DICOM specification of the VMAT treatment plan, which is preset as a default MLC leaf configuration used as an input to the VMAT treatment plan optimization process.

[0109] Fig.16An exemplary treatment plan adaptation and optimization process 800 is shown in FIG. 810 . To generate an adaptive plan 810 , a propagated synthetic (simulated) image (not shown) of the treatment link structure (i.e., target structure, OAR) and / or the patient is used as input to a plan generation algorithm S801 to generate an optimized plan adapted to the current treatment link. Typically, a synthetic image is required in order to generate a treatment plan from a stable and reproducible patient position. Generating a synthetic image involves a series of steps including: the treatment link image must be registered with the planning image including the planning structure; using the result of the registration as input to a deformable registration algorithm; and using the output of the deformable registration algorithm and data about the body contour from the treatment link image as input to the synthetic image generation algorithm.

[0110] like Fig.16 As shown in , the automatic plan generation algorithm of S801 takes the propagated treatment link structure (i.e., target structure, OAR) as input to modify the preliminary treatment plan 420 based on the new anatomical structure. Based on the result 711 of the similarity determination 700, the initial MLC leaf configuration 430 of the preliminary treatment plan 420 or the standard MLC leaf configuration 440 for one or more control points (CP) can be used to automatically generate a set of optimization parameters / criteria for adaptive plan generation in S801 in S802.

[0111] For example, when the calculated similarity coefficient (SC) is greater than the corresponding predetermined threshold (c), the MLC leaf configuration of each control point (CP) in the preliminary treatment plan 420 is used as an input metric for the optimization algorithm in S802. Then, a dose distribution for the current plan can be calculated in S803. The calculated dose distribution indicates the actual dose distribution in the target structure and OAR that can be achieved based on the current plan.

[0112] Since the mapping of the priority value P (i.e., priority 1, priority 2, priority 3, etc.) relative to the original clinical goal (CG) to the achieved value of the goal (i.e., the actual dose achieved for the goal) gives the plan quality for the goal, the plan quality for the current plan relative to the original clinical goal (CG) can be determined based on the calculated dose distribution. This can be done by determining the priority value P corresponding to the actual dose value for the goal (802). In the case where the quality goal metric is represented by a quality function (Q function), the P value can be read from the Q function generated for the clinical goal. The larger the value of the priority P of the goal, the better the goal is met, and therefore the better the quality of the plan for the goal.

[0113] Therefore, according to the calculated dose distribution of S803, a set of plan quality values ​​Ps for the current plan can be obtained (802), each plan quality value Ps being used for each original clinical purpose CG in the original clinical purpose CG.

[0114] Optionally, in S804, for example, the overall plan quality for the current plan may be obtained from the individual current plan qualities by taking the minimum value of all individual objective quality values.

[0115] Alternatively, the overall plan quality for the current plan may be obtained by defining the overall plan quality as the overall plan quality that minimizes the difference from the individual plan qualities.

[0116] Based on the value P of the overall quality of the current plan calculated in step S804, a decision can be made in S805 as to whether the current plan is acceptable (803), in which case the current plan becomes an adapted plan 810. In the event that the current plan is not acceptable, the control points used as input to the optimization algorithm in S802 are automatically adjusted via multi-resolution adjustment in S806. Adjusting the control points changes the dose distribution in S803, which in turn changes the plan quality value of the current plan. The overall plan quality is then calculated again in S804 to determine whether the subsequently obtained plan is acceptable or whether further adjustment of the control points in S806 is required.

[0117] The multi-resolution adjustment in S806 allows for multi-level (eg, four-level) adjustment, from coarse adjustment at level 1 to fine adjustment at level 4. Level 4 adjustment allows for fine tuning of the dose distribution.

[0118] For each iteration of the optimization algorithm, a decision may be made as to whether the individual plan quality value or the overall plan quality value for the current plan is acceptable, for example by comparing the calculated quality value / values ​​with the plan quality value / overall plan quality value of the preliminary treatment plan in step S805. If the difference exceeds a predetermined threshold, the result is determined to be unacceptable, in which case the control points (CP) are gradually modified in S806. The amount of adjustment, and therefore the level of resolution, may be determined based on the calculated difference.

[0119] When accepted, the adapted plan 810 is optimized to have a dose distribution having similar dose characteristics as the dose distribution of the originally planned treatment plan, ie, the preliminary treatment plan 420. Therefore, the adapted plan 810 should meet the original clinical objectives to a similar degree as the original / preliminary treatment plan 420.

[0120] If the similarity coefficient (SC) determined in 711 is less than a predetermined threshold (c), the optimization process proceeds using the control point parameters associated with the standard MLC blade configuration 440 .

[0121] Fig.18 Shown with Fig.16The process of processing plan adaptation and optimization 800 shown in FIG. 8 is similar to the process plan adaptation and optimization process 800 shown in FIG. 8 , except that the predetermined process plan 420' is used as an input to the adaptation plan generation and optimization process. When the predetermined plan 420' is used, when it is determined that the similarity coefficient (SC) is higher than the corresponding threshold value (c), the MLC blade configuration 430' of the predetermined process plan 420' is used to automatically generate a set of optimization parameters / standards in S802 for use in the adaptive plan generation in S801. If the similarity coefficient (SC) is lower than the corresponding threshold value (c), the standard MLC blade configuration 440 for one or more control points (CP) is used to automatically generate a set of optimization parameters / standards in S802 for use in the adaptive plan generation in S801.

[0122] Fig.19 and Fig. 20 A situation is illustrated where, instead of using the initial MLC blade configuration corresponding to the preliminary processing plan 420 as input to the optimization algorithm S802, the MLC blade configurations of different processing plans may be used to obtain an adaptive processing plan 810. Fig.19 , two treatment plans are shown, treatment plan A and treatment plan B. Both treatment plans are associated with treatment plans generated for the same patient in a previous treatment session. For example, treatment plan A corresponds to a treatment plan generated for a treatment session patient model obtained in a treatment session (i.e., treatment session A) prior to the current treatment session (i.e., for example, one day ago). On the other hand, treatment plan B may be a treatment plan generated for a treatment session patient model obtained in a treatment session (i.e., treatment session B) prior to both the current treatment session and treatment session A (i.e., for example, two days ago).

[0123] To determine which treatment plan MLC configuration to use as input to the treatment optimization algorithm in S802, a similarity measure for each of the two previous treatment session patient models is first determined in 700. Treatment session patient model A includes a planned target structure propagated onto a treatment session CBCT image obtained, for example, one day ago (i.e., CBCT image A of the previous date). Using any of the similarity measure calculation processes described previously, the target structure on CBCT image A is compared with the target structure 411B, 412B on the current treatment session CBCT image 510 of the current treatment session patient model 610. Similarly, using any of the similarity measure calculation processes described previously, treatment session patient model B includes a planned target structure propagated onto a treatment session CBCT image obtained, for example, two days ago (i.e., CBCT image B of the previous date).

[0124] In 711, the similarity coefficient SC(A) 710A obtained for the treatment link patient model A is compared with the similarity coefficient SC(B) 710B obtained for the treatment link patient model B. Next, the similarity coefficient with the highest value between the two is compared with the corresponding threshold value (c), and if it is higher than the threshold value, the MLC configuration of the treatment plan (the MCL configuration (430A) of the treatment plan A or the MCL configuration (430B) of the treatment plan B) is used as an input to the optimization algorithm of the treatment plan adaptation and optimization process 800 to obtain an adapted treatment plan 810. Otherwise, the standard MLC blade configuration 440 is used as an input to the optimization process.

[0125] If there are multiple session patient models and associated treatment plans, a similarity measure is performed for each session patient model (i.e., the similarity measure is calculated based on all previous daily CBCT images), and the optimization is performed using the MLC leaf sequence in the highest scoring existing treatment plan as the MLC leaf configuration. In this way, the optimization process can take advantage of the previous optimization in cases where the daily CBCT images are similar enough to the previous plan cases. If none of the treatment plans scores above the threshold (c), the MLC leaf sequence will revert to the standard leaf sequence.

[0126] The preliminary treatment plan 410, the predetermined treatment plan 420' and the adapted treatment plan 810 may be applied to the dose calculation algorithm of the dose calculation module 402A to calculate the corresponding radiation doses to be applied to the target structure / volume of the treatment session according to the corresponding treatment plan. The preliminary dose matrix, the predetermined dose matrix and the adapted dose matrix thus generated may be sent to a display device together with the generated preliminary plan, the predetermined plan and the adapted plan and the preliminary isodose values, the predetermined isodose values ​​and the adapted isodose values ​​to be displayed to the RTT / MTA user so that the user can select 900 the appropriate plan to be delivered to the patient.

[0127] The RTT / MTA is also provided with one or more tools (hardware, software, etc.) to evaluate the predetermined plan and the adapted plan. The tools may include a tool that provides the isodose distribution of the preliminary plan on the planning image; provides the isodose distribution of the predetermined plan on the processing link image; provides the isodose distribution of the adapted plan on the processing link image; and dose volume histograms (DVH) of the preliminary plan, predetermined plan, and adapted plan. The tools may also include tools for clinical purpose evaluation to provide the clinical purpose value of the plan and the clinical actual value of the predetermined and adapted plans. Optionally, the RTT / MTA may be presented with other images, such as images from internal radiation treatment, which may assist in the plan selection process. The RTT / MTA may select which information to display and which additional tools to use to help select the most appropriate plan.

[0128] The RTT / MTA uses the planned clinical objectives and planned clinical objective values ​​in the instruction set to quantitatively evaluate whether the predetermined and / or adapted plan is medically necessary for the treatment session of the day (i.e., the current treatment session). To this end, the planned clinical objective values ​​and actual clinical values ​​for the predetermined and adapted treatment plans are presented to the RTT / MTA on a display, and the RTT / MTA selects the treatment plan that provides the clinical value that is closest to the planned clinical objective value.

[0129] After the user selects a treatment plan, radiation treatment will be performed according to the selected plan 900. The prescribing physician or any other qualified physician who generated the original / preliminary treatment plan 420, adaptive instructions 230, and adaptive workflow 220 can review and / or monitor treatment delivery offline, such as Figure 2 as shown in .

[0130] Throughout adaptive radiation treatment, DICOM artifacts (benchmark plans, images, structures, doses, etc.) can be used to verify the planned checks to ensure delivery on the treatment device, and the DICOM artifacts are provided for independent quality assurance. Optionally, the independent verification application transmits the results of the verification (QA) back to the adaptive workflow application.

[0131] It is therefore apparent that the disclosed subject matter enables the use of similarity metrics to determine optimization strategies in adaptive radiation therapy.

[0132] It is also apparent that the disclosed subject matter implements an automated process for an adaptive radiation therapy session to: obtain an instruction set comprising a plan quality value for a preliminary treatment plan for a planning patient model; generate a treatment session patient model using information from the instruction set; determine a similarity measure between a target structure of the planning patient model and a corresponding target structure of the treatment session patient model; and generate an adapted treatment plan for the treatment session patient model by: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan; the treatment plan generation algorithm comprising optimization parameters generated based on control points, the control points being automatically selected based on the determined similarity measure; and generating an adapted treatment plan by optimizing the current plan based on the plan quality value of the preliminary treatment plan.

[0133] Therefore, it is also apparent that the disclosed subject matter implements a method for adapting a preliminary radiation treatment plan, the preliminary radiation treatment plan being used for a current treatment session of an adaptive radiation therapy session, the method comprising: obtaining an instruction set, the instruction set comprising a plan quality value of the preliminary treatment plan for a patient; using the instruction set, performing a series of automated steps to generate a current treatment session patient model; determining a similarity measure between a target structure of the current treatment session patient model and corresponding target structures of a plurality of previously obtained treatment session patient models; determining a previously obtained treatment session patient model having a highest similarity measure value; and generating a current treatment session patient model by: An adapted treatment plan for a patient model: generating a current plan using a preliminary treatment plan as input to a treatment plan generation algorithm; the treatment plan generation algorithm comprising optimization parameters generated based on control points, the control points being automatically selected based on a determined similarity metric; and generating an adapted treatment plan by optimizing the current plan based on a plan quality value of the preliminary treatment plan, wherein the control points comprise an MLC blade configuration, and wherein when a highest similarity metric value is below a predetermined threshold, a standard MLC blade configuration is selected, and when a highest similarity metric value is above a predetermined threshold, an MLC blade configuration of a treatment plan of a patient model of a previous treatment link having the highest similarity metric value is selected.

[0134] It is also apparent that the disclosed subject matter implements a system for implementing an automated workflow for an adaptive radiotherapy session for a patient, comprising: a computer processing system configured to: obtain an instruction set comprising a plan quality value for a preliminary treatment plan for the patient; and, using the instruction set, perform a series of automated steps to: generate a current treatment session patient model; and generate an adapted treatment plan for the treatment session patient model by: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan, optimizing the current plan based on optimization parameters, the optimization parameters being generated based on control points, the control points being selected based on a similarity metric determined between a target structure of the current treatment session patient model and a corresponding target structure of a planning patient model or one or more previously obtained treatment session patient models, wherein the control points comprise an MLC leaf configuration, and wherein when the highest similarity metric value is below a predetermined threshold, a standard MLC leaf configuration is selected, and when the highest similarity metric value is above a predetermined threshold, the MLC leaf configuration of the treatment plan of the previous treatment session patient model having the highest similarity metric value is selected.

[0135] It should be understood that aspects of the disclosed subject matter may be implemented in whole or in part in hardware, hardware programmed by software, software instructions stored on a computer-readable medium (eg, a non-transitory computer-readable medium), or any combination of the above.

[0136] For example, components of the disclosed subject matter (including components such as controllers, processes, or any other features) may include, but are not limited to, a personal computer or workstation or other such computing system that includes a processor, microprocessor, microcontroller device, or is comprised of control logic including an integrated circuit (such as, for example, an application specific integrated circuit (ASIC)).

[0137] Features discussed herein may be performed on a single or distributed processor (single core and / or multi-core), by components distributed across multiple computers or systems, or by components co-located in a single processor or system. For example, aspects of the disclosed subject matter may be implemented via a programmed general purpose computer, an integrated circuit device (e.g., an ASIC), a digital signal processor (DSP), an electronic device programmed with microcode (e.g., a microprocessor or microcontroller), a hardwired electronic or logic circuit, a programmable logic circuit (e.g., a programmable logic device (PLD), a programmable logic array (PLA), a field programmable gate array (FPGA), a programmable array logic (PAL)), software stored on a computer readable medium or signal, an optical computing device, a networked system of electronic and / or optical devices, a special purpose computing device, a semiconductor chip, a software module or object stored on a computer readable medium or signal.

[0138] When implemented in software, the function can be stored on a computer-readable medium or transmitted via a computer-readable medium as one or more instructions or codes. The steps of the algorithm or method disclosed herein can be embodied in a processor executable software module, which can reside on a computer-readable medium. Instructions can be compiled from source code instructions provided according to a programming language. The programming instruction sequence and the data associated therewith can be stored in a computer-readable medium (e.g., a non-transient computer-readable medium) such as a computer memory or storage device, which can be any suitable memory device, such as, but not limited to, a read-only memory (ROM), a programmable read-only memory (PROM), an electrically erasable programmable read-only memory (EEPROM), a random access memory (RAM), a flash memory, a disk drive, etc.

[0139] As used herein, computer-readable media include computer storage media and communication media, including any medium that is convenient for transferring a computer program from one place to another. Therefore, the storage medium can be any available medium that can be accessed by a computer. As an example and not limitation, such computer-readable media can include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage devices or any other medium that can be used to carry or store the required program code in the form of an instruction or data structure and can be accessed by a computer.

[0140] Moreover, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a website, server or other remote source using a transmission medium (e.g., coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio and microwave), then the transmission medium is included in the definition of computer-readable medium. In addition, the operations of a method or algorithm may reside as one (or any combination thereof) of a set of codes and / or instructions on a machine-readable medium and / or computer-readable medium, which may be incorporated into a computer program product.

[0141] Those skilled in the art will readily appreciate that the above description is not exhaustive, and that various aspects of the disclosed subject matter may be implemented in ways other than those specifically disclosed above. In fact, embodiments of the disclosed subject matter may be implemented in hardware and / or software using any known or later developed system, structure, device, and / or software by those skilled in the art according to the functional description provided herein.

[0142] In this application, unless specifically stated otherwise, the use of the singular includes the plural, and the individual use of "or" and "and" includes the other, i.e., "and / or." In addition, the use of the terms "include" or "have" and other forms are intended to have the same effect as "comprising" and therefore should not be construed as limiting.

[0143] Any range described herein should be understood to include the endpoints and all values ​​between the endpoints. Whenever "substantially," "about," "substantially," "close to," or similar language is used in combination with a particular value, unless expressly stated otherwise, a variation of up to and including 10% of that value is intended.

[0144] The terms "system," "device," and "module" are used interchangeably herein, and the use of one term in the description of an embodiment does not preclude the application of the other terms to that or any other embodiment.

[0145] The present disclosure allows for many substitutions, modifications, and variations. Although specific examples are shown and described in detail for the purpose of illustrating the application of the principles of the present invention, it should be understood that the present invention may be embodied in other ways without departing from these principles. For example, the disclosed features may be combined, rearranged, omitted, etc., to produce additional embodiments, and certain disclosed features may sometimes be advantageously used without corresponding use of other features. Therefore, the applicant intends to cover all such substitutions, modifications, equivalents, and variations within the spirit and scope of the present invention.

Claims

1. An automated method for an adaptive radiation therapy session, comprising: obtaining an instruction set including a plan quality value for a preliminary treatment plan for a planning patient model; generating a treatment phase patient model using information from the set of instructions; determining a similarity measure between a target structure of the planning patient model and a corresponding target structure of the treatment session patient model; as well as Generate an adapted treatment plan for the patient model of the treatment link by performing the following operations: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan; The treatment plan generation algorithm includes optimization parameters generated based on control points, the control points automatically selected based on the determined similarity measure; as well as The adapted treatment plan is generated by optimizing the current plan based on the plan quality value of the preliminary treatment plan. 2 . The method of claim 1 , wherein the target structures include one or more contours of a primary tumor and contours of one or more organs at risk (OARs).

3. The method of claim 2, wherein the instruction set includes information about: a planned radiation dose, a planned clinical objective, a planned clinical objective value, a planned patient model, a list of influencer structures, and control points.

4. The method of claim 3, wherein generating the treatment phase patient model comprises: generating a processed segment image of a portion of the patient including an anatomical structure of interest; as well as The target structure of the planning image is propagated to the processing link image. The method of claim 4 , wherein the propagation of the target structure is guided by one or more influencer structures from the list of influencer structures.

6. The method of claim 1, wherein the similarity metric is one or a combination of the following: Sorensen-Dice coefficient, Jaccard index, or Hausdorf metric.

7. The method of claim 6, wherein the determining of the similarity measure comprises: Determination of the amount of overlap between corresponding target structures.

8. The method according to claim 7, further comprising: Prior to the determination of the similarity measure, a rigid registration is applied between the planning patient model and the treatment session patient model.

9. The method of claim 8, wherein the rigid registration comprises: obtaining a planned isocenter position for the planning patient model from the set of instructions; Determining a central location for collecting a patient model for the processing step; aligning the target structure in the treatment session patient model with the corresponding planning target structure in the planning patient model; determining a difference between a position of the planning isocenter and a position of the acquisition isocenter; as well as The processing link isocenter position is determined by applying the determined difference to the acquisition isocenter position.

10. The method of claim 6, wherein the control point comprises an MLC blade configuration. The method of claim 10 , wherein when the similarity metric is below a predetermined threshold, a standard MLC blade configuration is selected.

12. The method of claim 11, wherein the MLC blade configuration of the preliminary process plan is selected when the determined similarity metric is above the predetermined threshold.

13. The method of claim 12, wherein the processing link isocentral location is used as input to the plan generation and optimization algorithm.

14. The method of claim 1, wherein the planning image is a CT image or a CBCT image obtained in a previous processing session, and the processing session image is a current CBCT image.

15. The method according to claim 1, further comprising: Determining a dose distribution for the current plan based on the selected control points; determining a plan quality value for the current plan based on the determined dose distribution; as well as comparing the determined plan quality value with the preliminary plan quality value, Based on the result of the comparison, the control points for optimizing the current plan are gradually adjusted.

16. A method for adapting a preliminary radiation treatment plan for use in a current treatment session of an adaptive radiation therapy session, comprising: obtaining an instruction set including a plan quality value for a preliminary treatment plan for a patient; Using the set of instructions, a series of automated steps are performed to generate a patient model of the current treatment link; determining a similarity measure between a target structure of the current treatment session patient model and corresponding target structures of a plurality of previously obtained treatment session patient models; determining the previously obtained treatment session patient model having the highest similarity measure value; and Generate an adapted treatment plan for the patient model of the current treatment link by performing the following operations: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan; The treatment plan generation algorithm includes optimization parameters generated based on control points, the control points automatically selected based on the determined similarity measure; as well as generating the adapted treatment plan by optimizing the current plan based on the plan quality value of the preliminary treatment plan, wherein the control point comprises an MLC blade configuration, and wherein when the highest similarity metric value is below a predetermined threshold, a standard MLC blade configuration is selected, and When the highest similarity metric value is above the predetermined threshold, the MLC blade configuration of the treatment plan of the previous treatment session patient model having the highest similarity metric value is selected.

17. The method of claim 16, wherein the similarity metric is one or a combination of the following: Sorensen-Dice coefficient, Jaccard index, or Hausdorf metric.

18. The method of claim 17, wherein the target structures include one or more outlines of a primary tumor and outlines of one or more organs at risk (OARs).

19. The method of claim 18, wherein the set of instructions includes information regarding planned radiation doses, planned clinical objectives, planned clinical objective values, planned patient models, influencer structure lists, and control points.

20. The method according to claim 19, further comprising: Determining a dose distribution for the current plan based on the selected control points; determining a plan quality value for the current plan based on the determined dose distribution; as well as comparing the determined plan quality value with the preliminary plan quality value, Based on the result of the comparison, the control points for optimizing the current plan are gradually adjusted.

21. A system for implementing an automated workflow for an adaptive radiation therapy session for a patient, comprising: A computer processing system configured to: obtaining an instruction set, the instruction set comprising a plan quality value for a preliminary treatment plan for the patient; as well as Using the described set of instructions, a series of automated steps are performed to: Generate a patient model of the current treatment link; as well as Generate an adapted treatment plan for the patient model of the treatment link by performing the following operations: using the preliminary treatment plan as input to a treatment plan generation algorithm to generate a current plan, optimizing the current plan based on optimization parameters generated based on control points selected based on similarity measures determined between target structures of the current session patient model and corresponding target structures of a planning patient model or one or more previously acquired session patient models, wherein the control point comprises an MLC blade configuration, and wherein when the highest similarity metric value is below a predetermined threshold, a standard MLC blade configuration is selected, and When the highest similarity metric value is above the predetermined threshold, the MLC blade configuration of the treatment plan of the previous treatment session patient model having the highest similarity metric value is selected.

Images