Systems and methods for automated radiation therapy treatment planning - Patents.com

Abstract

A method for generating a radiation therapy treatment plan using a multi-leaf collimator (MLC), a gantry, and a couch, providing a number of candidate arc paths with isocenters and a maximum number of arc beams for the treatment plan and a maximum number of target groups according to the maximum number of arc beams, providing a shape and position of each target in a target set, calculating a target segment among possible target segments of the target set, the possible target segments including a maximum number of target groups, each of the possible target segments including a target group, determining a cost of each target group of the possible target segments taking into account the candidate arc paths, the gantry, and the MLC angles, determining a cost of the possible target segments and current candidate arc paths by summing the costs of the target groups of the possible target segments, repeating the calculation, determination, and summation for the possible target segments and candidate arc paths, and selecting an optimal target segment and candidate arc path with a minimum total treatment plan cost.

Description

[Technical field]

[0001] Technical Field The present invention relates to systems and methods for radiation therapy treatment planning, in particular the present invention relates to systems and methods for automated stereotactic radiosurgery planning for multiple metastases (meta) and also to treatment methods thereof. [Background technology]

[0002] background

[0002] Radiotherapy treatment is known to use beams of photons, electrons, protons or other ions and direct them towards a treatment area of ​​a patient to treat that area or volume. The shape of the beam is usually selected so that the treatment area receives the desired dose while limiting the dose to the surrounding tissue. In particular, sensitive organs known to be at risk must be protected as much as possible. To achieve this, a gantry that can rotate around the patient is often used to provide radiation beams from different angles so that all beams reach the target while each part of the surrounding tissue receives the dose only from one or several beams, or in the case of arc beams, to a lesser extent than the target. The gantry can rotate completely around the patient or partially along a part of the circumference. The gantry can provide radiation in a continuous arc as it moves around the patient, or it can stop at an angle to deliver a static beam. The fractionated doses in the treatment can consist of multiple arcs or static beams or a combination of both. The patient couch on which the patient is placed and secured may also rotate during delivery of radiation or between static or arc beams to modify the irradiation direction relative to the patient. There are other means for changing the beam direction, for example the radiation source may be mounted on a mobile robotic arm, but for the purposes of this discussion a gantry is used as an example. Typically, radiation therapy such as photon therapy uses a collimator located in the beam plane, i.e. perpendicular to the central beam axis, to shape the beam so that the deposited dose matches the prescribed dose as accurately as possible. Typically the collimator is a multi-leaf collimator (MLC) as described later in this specification.

[0003]

[0003] In the presence of multiple metastases (multi-metastases), radiosurgery planning becomes exponentially more difficult when aiming to treat such multiple metastases simultaneously in one treatment or treatment session using a limited number of beams in order to save treatment time. In general, saving treatment time in radiotherapy has several advantages, such as saving time and therefore cost, reducing patient discomfort, and reducing the time that the patient can potentially move and displace the target area, which is the tumor, from the target volume.

[0004]

[0004] In conventional multi-meta radiosurgery planning, the treatment planner usually manually plans and also groups multiple targets (metastases) by considering several parameters. This task is time-consuming and quite difficult because the number of potential target groups increases exponentially with the increase in the number of metastases. Manual target grouping is very challenging because of many unknown variable parameters such as collimator angle, couch angle and gantry angle, so it is uncertain whether optimal grouping will be achieved. In addition to these parameters, reduction and avoidance of non-target exposure area (ENTA) must be considered during planning. As described herein with reference to the figures, various metastasis shapes or metastasis groups can lead to ENTA problems. Therefore, optimizing the plan is a heavy burden for the treatment planner. In light of these challenges, efforts have been made to automate the planning of multi-meta cases.

[0005]

[0005] Although some automated tools and solutions exist, they only address a limited number of parameters, such as specific MLC angles and / or gantry angles, and only examine a limited amount of target groups of the complete target set, and do not consider a given number or all possible groups of targets.

[0006]

[0006] In addition, some of the known tools and solutions have difficulty in delivering equal dose or dose target to each target of the target set due to jagged MLC window or non-exposure, and there is a risk of non-conformal dose delivery. Furthermore, when multiple metastases at different depths are treated simultaneously by the beam, they are likely to receive different doses due to the relationship of the depth-dose curve of the photon beam. In radiotherapy, each target, metastasis or tumor should receive a radiation dose that can be planned at the target, if possible, which may mean that the dose is uniform, and may mean that the dose distribution is not uniform across the target (e.g., the center of the target has a higher dose), but the delivered dose should be as close as possible to the dose target. This is not always possible with the methods and solutions of the prior art.

[0007] Another problem with prior art solutions, such as intensity modulated arc therapy (VMAT), is that they allow for jagged apertures in the multi-leaf collimator. Such non-conformal apertures (see Figs. 7a and 7b) can lead to problems when dose levels in the target volume need to be controlled and uniformly distributed. In general, jagged MLC apertures should be avoided. Jagged MLC apertures are MLC apertures where a particular leaf extends significantly further than its neighboring leaves. With jagged leaf apertures and a continuously moving gantry, providing robust treatments is very challenging and creates problems in providing quality assurance. In the method proposed herein, jagged MLC apertures can be avoided by using conformal systems and methods, particularly by using conformal arc beams. However, it should be noted that jagged MLC apertures may be advantageous in some cases, as they may provide additional degrees of freedom for modulation to help reach the target dose goal.

[0008]

[0008] In addition, when the objective is to create a VMAT plan, the process of optimizing such a VMAT plan is greatly simplified by starting the optimization from a high-quality conformal arc plan. Starting from a high-quality conformal arc plan typically results in a final arc plan with higher quality than would otherwise be obtained. Thus, the present invention also provides a method for improving VMAT plans in stereotactic radiosurgery. Summary of the Invention [Problem to be solved by the invention]

[0009] overview

[0009] In light of the above, the object of the present invention is to provide a method that makes it possible to automate planning of multi-meta cases and improve the efficiency of multi-meta treatment using stereotactic radiosurgery (SRS).

[0010]

[0010] A further object is to provide a method that enables optimal treatment to be provided that can provide dose targets across the entire target, even when a limited number of beams are required to treat a set of targets in the same treatment or treatment session.

[0011]

[0011] A further object is to provide a method which saves time and allows optimizing multi-meta radiation therapy. [Means for solving the problem]

[0012]

[0012] In a first aspect of the present invention, the inventors of the present invention have realized that the planning process for radiation treatment of multiple metastases can be improved by automating target grouping. Thereby, a set of targets to be treated simultaneously is divided into several target groups by considering possible pairs of gantry and collimator angles, analyzing some or all possible target sections of the target set, and thereby also analyzing some or all possible combinations of target groups, with the consistent objective of avoiding bridge gap opening (the term bridge gap opening will be explained later in this specification with reference to Figs. 5a and 5b) and / or avoiding exposure of radiation to non-target exposed areas (ENTA). In short, bridge gap opening is a configuration of a multi-leaf collimator (MLC) when the leaves cannot be fitted to two or more targets without exposing non-target tissue or non-target areas between the targets.

[0013]

[0013] In a second aspect of the present invention, the inventors have realized that if a target set exceeds a certain amount N of targets or if a target is located too far from another target, a clustering step can be performed to prepare the target set for optimal grouping automation. Such a clustering step can include the steps of selecting a maximum number M of clusters based on the number of targets, limiting the number of targets in a cluster to N or less, and selecting targets in a cluster based on the maximum allowable distance between two targets of the same cluster. The second aspect requires an input of a target set including the shape of the targets and their relative positions to each other, and then an optimal clustering is provided as an output, taking into account the above-mentioned parameters such as possible pairs of gantry and collimator angles, as well as geometric parameters such as the maximum allowable distance between two targets of the same cluster.

[0014]

[0014] In a third aspect, the invention relates to a solution for performing dose control shielding at the input of an arc plan with calculated dose and dose level targets for each target. Once the dose level targets for each target are calculated or determined, targets with excessive dose levels are shielded over a part of the arc path or arc beam via MLC or at some control points. This means that for at least some arc beams along a candidate arc path, the leaves of the MLC are completely closed over the target with excessive dose level. Optionally, the monitor unit (MU) level of the beam is changed. According to the method according to the third aspect, it is possible to modify the start and end gantry angles to avoid the presence of segments or arc beams with completely closed MLC. As an output, the method according to the third aspect provides a modified, preferably conformal, arc plan that improves the achievement of dose level targets for all targets while at the same time taking advantage of the efficient and convenient MLC opening properties of the conformal arc plan.

[0015]

[0015] Disclosed herein is a treatment planning method for generating a radiation therapy treatment plan in which a target set is to be treated. The method uses a multi-leaf collimator (MLC) for shaping an arc beam, a gantry including the MLC, which can rotate at least partially around a patient, and a couch, chair or other structure intended to position the patient. The MLC is rotatable about a beam axis, thereby defining a collimator angle, the gantry is rotatable about a gantry axis, thereby defining a gantry angle, and the couch is rotatable about a z-axis, thereby defining a couch angle. The method includes: - providing a number of candidate arc paths having an isocenter, a maximum number of arc beams for the treatment plan, and a maximum number of target groups depending on the maximum number of arc beams; - providing a three-dimensional shape and position of each target of a target set; - calculating target partitions of at least some possible target partitions of a target set based on a current candidate arc path, the at least some possible target partitions including a maximum number of target groups, each of the at least some possible target partitions including a target group, whereby optionally the targets are each part of at least one target group; - for each target group of possible target sections, determining a cost taking into account a current candidate arc path, at least one gantry angle and at least one MLC angle; - determining the cost of the possible target partition and the current candidate arc path by summing the cost of each target group of the possible target partition; - repeating the steps of calculating and determining for each of at least some of the target partitions and candidate arc paths; - selecting an optimal target segment and arc path that minimizes the total cost of the treatment plan; Includes.

[0016]

[0016] It should be noted that it is not always necessary for each target to be part of at least one target group or at least one target compartment, but in some cases, such as the examples discussed herein, target compartments that meet this requirement can be used and configured for treatment.

[0017]

[0017] According to the above, it is not necessary to perform calculations, determinations and sums for each possible target partition, even if it were possible. In order to save resources, such as computational resources, only relevant target partitions can be considered. Such a selection of target partitions can be obtained or provided in another way, for example, by dynamic programming, whereby the selection of target partitions used is known to include the optimal target partitions. For example, if it is known that a particular target group leads to high costs, the target partitions that include that group can be removed from the set of target partitions considered.

[0018]

[0018] The above-mentioned method allows a target set to be provided to a system, such as a computer or radiation treatment planning system, so that the method automatically calculates at least some possible target segments for evaluation and finds the optimal target segments for treatment. The target segments thereby include some target groups, and as can be understood, the target set may have many target segments. The target segments and target groups may have the condition that the targets are only part of one group each, or only part of one group each per candidate arc path, in order to avoid some targets being treated more than others, which would result in too high a radiation dose level. The automation of target segmentation and therefore target grouping relieves the treatment planner from very complex and time-consuming planning. The treatment method also improves treatment results, since it considers all possible gantry angles, MLC angles, total costs (e.g., total ENTA costs or other costs described below) and target segments and target groups. Specific limitations for the calculations may be provided to the system, such as the maximum number of targets in a target group, the maximum number of target groups in a target segment, and potentially geometric considerations. Geometric considerations may, for example, specify that all targets of a target group must fit into a predefined volume so that they are geometrically close to one another, or that targets of the same target group must not be closer than a predefined distance to one another. Alternatively, it may specify that the projection of all targets of a target group onto a plane perpendicular to the central beam axis must fit into a predefined area. Geometric considerations or conditions may relate to maximum and / or minimum allowable distances and / or volumes.

[0019]

[0019] The calculation step can consider all possible target partitions. Alternatively, the calculation step can consider only target partitions that have less than a specified number of target groups. Alternatively, the calculation step does not need to include target partitions that have target groups that are known to lead to poor solutions in the target partitions that are considered. Alternatively, a maximum specified number of target partitions can be considered.

[0020]

[0020] For optimization purposes, all the candidate arc paths provided can be used.

[0021] Alternatively, a static field or static beam can be considered for the candidate arc paths.

[0022]

[0022] In an embodiment, known non-optimal target sections, such as, for example, target sections including all targets within one target group or target sections including at least two targets positioned far apart or very close to each other (such as more than 5-6 cm apart from each other or very close, such as a few mm (millimeters)), can be directly excluded from the optimization method described above.

[0023] In an embodiment, the calculating step may further take into account at least one couch angle.

[0024]

[0024] By taking into account at least one couch angle, the optimal target compartment result can be further improved since the couch angle is an additional parameter that can be adjusted.

[0025] In an embodiment, the maximum number of target groups is provided by the treatment planner or is any predetermined number between 1 and 6 per candidate arc beam.

[0026]

[0026] The maximum number of target groups may be based on the maximum number of arc beams allowed or vice versa.

[0027]

[0027] In another embodiment, the calculation step may further include dividing the arc path into control points spaced at regular intervals, each interval corresponding to a gantry angle segment, and at each control point, an optimal MLC angle and / or an optimal couch angle is calculated. During this calculation step, movement constraints such as maximum rotation speed of the MLC or maximum angle rotation of the MLC may be taken into account.

[0028]

[0028] To further optimize the target division and select the optimal target group, predefined control points can be used, typically spaced at regular intervals along the arc path, and therefore along the gantry path. In addition, several MLC angles and couch angles can be considered at each control point. The control points can be arranged, for example, at gantry angle intervals of 1°, 2° or 4° along the arc path. The distance of the control points and therefore the distance of the regular intervals can be selected based on the available computing power of the current computer.

[0029]

[0029] In a further embodiment, for each control point, an optimal MLC opening is calculated, whereby the optimal MLC opening is selected based on the minimum cost and the optimal MLC angle relative to the previous MLC angle and the next MLC angle at the previous control point and the next control point.

[0030]

[0030] As above, if the MLC angle needs to be changed between control points, further cost is added.

[0031]

[0031] In some methods, changing the MLC angle from one control point to the next control point may lead to additional cost at a point, for example, because adjusting the MLC angle between control points may be time-consuming and very difficult. However, if a bridge gap is open at a particular MLC opening at the current control point or another ENTA exists, the cost for such a situation may be high, and accordingly, the method may choose to rotate the MLC and provide another MLC angle to avoid the ENTA cost or another additional cost.

[0032]

[0032] In light of the above, according to one embodiment of the present specification, the calculation step may further include dividing the arc path into control points spaced at regular intervals, each interval corresponding to a gantry angle segment, the MLC angle being maintained at a fixed position for each control point along the candidate arc path, and the cost of each candidate MLC angle being calculated for each control point to find an optimal fixed MLC angle for each candidate arc path.

[0033]

[0033] In another embodiment, costs may be added if the couch angle needs to be changed between nearby control points.

[0034]

[0034] Changing the couch angle can cause delays or uncertainty, which can add cost if such a change is necessary.

[0035]

[0035] In a further embodiment, the maximum number of arc beams corresponds to the maximum number of target groups.

[0036]

[0036] The above can facilitate achieving correct dose level targets for each target and target group, which can further improve and shorten patient treatment times.

[0037] In another embodiment, the arc beam may be a conformal arc beam.

[0038]

[0038] The use of conformal arc beams results in a clear and understandable dose level distribution in that each target receives radiation using the efficient and useful MLC aperture properties of the conformal arc plan, or receives no radiation delivery at all if the leaves completely cover the target. The conformal MLC aperture is intended to be either completely open or completely closed over the entire target, but may accidentally open, for example when it needs to move over the target to be able to reach the desired position of the MLC at the next control point. Speed ​​limitations may be considered during such movements. Jagged leaf positions, where some leaves extend significantly further into the aperture than neighboring leaves, may be avoided. However, according to the third aspect of the present disclosure, complete dose shielding may be tolerated, and thus, at a control point, some leaves of the MLC may be adjusted so that one or some targets do not receive any radiation at all, to provide a dose level that matches or at least very close to the dose level target.

[0039]

[0039] In a further embodiment, the method may further include obtaining, for each target group of the optimal target sections, a minimum-cost arc beam having a minimum-cost MLC angle and gantry angle pair, typically by respecting a minimum gantry angle spacing, and delivering these minimum-cost arc beams, whereby each arc beam is matched to one target group of the optimal target sections such that each target group receives one arc beam for treatment.

[0040]

[0040] To avoid selecting a gantry spacing that is too small, a minimum gantry angle spacing can be provided.

[0041]

[0041] For each target group of the optimal target section, a minimum-cost arc beam or arc path with a minimum-cost pair of MLC and gantry angles may be provided to generate a treatment possibility for the current multi-meta case. Such a minimum-cost arc beam may be close to the candidate arc path provided initially, but the angles may be slightly different.

[0042] The methods and / or steps described can be used as a starting point for VMAT optimization. Thus, if a treatment planner has a multi-metastatic, and therefore multi-meta, case at hand, the treatment planner can use the methods described herein to optimize the starting point for other treatment methods, such as VMAT.

[0043]

[0043] In another embodiment, at least some possible target sections are considered for various arc paths, and an arc plan including arc beams and arc paths is provided based on the optimal target section and least cost arc beam, taking into account limitations on the maximum number of arc beams, such as a maximum number or a maximum number of target groups in total for an arc beam.

[0044]

[0044] Thus, the method can simultaneously consider several arc paths for at least some possible target segments to find the best treatment plan.

[0045]

[0045] In another embodiment, the method further comprises: - receiving an arc plan and a calculated dose for each target in an optimal target section, the arc plan may include several arc paths; - providing a dose level target for each target and adjusting the arc plan so that each target achieves at least the dose level target; - achieving or providing the correct dose level target for each target instead of using target shielding to achieve the correct relationship between the targets and then scaling up or optimizing the control point monitor unit MU in the next step; - identifying targets that achieve a dose level that exceeds a dose level goal; - calculating, for all targets in the optimal target section, a required dose control shielding in which the leaves of the MLC are fully closed for at least a portion of some arc beams for targets whose dose levels exceed the dose level target in order to optimize the dose delivery and thus the dose level target; It may further include.

[0046]

[0046] The method of the previous paragraph mentioned above can be used independently of any other method disclosed herein. The method mentioned above can be called a dose control shielding method. This dose control shielding method, as mentioned above, is the third aspect of the present invention disclosed herein, and stands alone independently of other method steps disclosed herein. At the same time, the dose control shielding method can be used and combined with any method step or feature disclosed herein, independently of other features and method steps.

[0047]

[0047] Any arc plan that can be calculated or provided by the radiation therapy planning method can be used as an input method for the dose control shielding method disclosed herein to achieve optimal target dose levels in the target group on the arc beam. In addition, conformal arc plans, in particular, can be used as input for the dose control shielding method according to the previous paragraph.

[0048]

[0048] The output of the dose control shielding method can be a modified conformal arc plan, whereby at some positions or control points, a part of the target is shielded from radiation through the MLC. Typically, the MLC is open to at least one target at each control point. In some cases, at some control points, the MLC can be fully closed. By modifying the start and end gantry angles, and potentially also the MLC angle, control points or beams with fully closed collimators can be avoided, because such closed segments are a waste of time during treatment. Another option is to ignore fully closed collimator or MLC segments and proceed on such segments without applying radiation to these segments.

[0049] As explained, the control points and / or starting and ending gantry angles may be selected such that there are no completely closed control points or segments where all targets are occluded.

[0050]

[0050] By modifying the start and end gantry angles, it is possible to accommodate and / or further include segments that are ignored because in these segments the MLC aperture is fully closed. Alternatively, segments where the start and end gantry angles are not modified but the MLC configuration or aperture is fully closed can be ignored.

[0051] In another embodiment according to the disclosure of this specification, according to the second aspect, a clustering step is performed before the step disclosed in the first aspect when the distance between two targets exceeds a certain threshold or the number of targets exceeds N targets. The clustering step may include the steps of selecting a maximum number M of clusters based on the number of current targets, limiting the number of targets in a cluster to N or less, and selecting targets in a cluster based on the maximum allowable distance between two targets of the same cluster or geometric considerations or conditions.

[0052]

[0052] As input, the position and three-dimensional shape of the target may be provided.

[0053] In another embodiment, if the total number of target groups in the current target section exceeds a certain threshold, such as 15-20 target groups, additional costs may be added.

[0054]

[0054] This may lead to improved targeting and optimisation of the planning process.

[0055]

[0055] In another embodiment, the clustering step can directly incorporate geometric considerations as initial conditions. Thus, if it is detected or determined that two targets are too far from each other and thus exceed a distance threshold (threshold may be 6 cm), the clustering step can be directly started without any restriction on the number of current targets N. Any combination of the maximum number of targets N, the maximum number of clusters M and geometric considerations or geometric conditions can be used to start the clustering step.

[0056]

[0056] The clustering step may be used when there are more than 3 targets, more than 4 targets and / or more than 5 targets, preferably when there are more than 10-15 targets. The clustering step reduces the complexity in that the number of targets treated with a set of arcs or methods is reduced and they are rather geometrically located closer to each other. Instead, a maximum allowable distance between any two targets of the same cluster may be calculated, as mentioned above, for example in cm. 3 A volume can be given in units and all targets of a cluster should fit within this volume for ease of treatment and targeting. These geometric conditions, such as maximum volume or maximum allowable distance, are called geometric conditions or geometric considerations.

[0057]

[0057] The clustering step mentioned above is the second aspect according to the disclosure of this specification, and can be used independently of any other method or feature disclosed herein. The clustering step or the clustering method is not related to the first or third aspects of the present invention disclosed herein. The clustering step or the method can be used before conventional radiation therapy such as IMRT or VMAT to prepare and reduce the number of targets treated with the exact same treatment to a manageable level.

[0058]

[0058] As an example, an example is provided below to better illustrate the clustering method and steps. A brain tumor patient has seven metastases in his head, some of which are near the right ear and others are near the left eye. The maximum allowable distance between targets (metastases) is 6 cm. Thus, the requirement is that all targets in a cluster must be spaced no more than 6 cm apart from each other. The maximum number of targets in one cluster is set to 4, which corresponds to N, so that the total number of targets, 7, exceeds the number of targets allowed in one cluster, 4. The maximum number of clusters M in this case is 3. As long as the criterion of the maximum distance being 6 cm is met, the first cluster (e.g., the cluster close to the right ear) contains four targets, and the second cluster (e.g., the cluster close to the left eye) contains the remaining three targets. Theoretically, a third cluster is possible, but is not necessary in order not to exceed the maximum number N of four targets per cluster and not to violate the condition of the maximum allowable distance of 6 cm. Once the clusters are defined, the methods according to the first and third aspects or any other known radiation therapy planning can be performed.

[0059]

[0059] In the above example, it should be noted that alternatively, there may be a cluster with only one target, and the other two clusters contain the remaining targets. The method evaluates which clustering is optimal, taking into account provided constraints such as the maximum number of targets in a cluster and the maximum number of clusters. These constraints may depend on other circumstances, such as the available arc beams or treatment time limits specified by the treatment planner.

[0060]

[0060] Also disclosed herein is a computer program product comprising computer readable means which, when executed on a computer, causes the computer to perform any of the aspects or method steps disclosed herein.

[0061]

[0061] Also disclosed herein is a non-transitory computer readable medium encoded with computer executable instructions that, when executed on a computing device, cause the device to perform method steps or aspects of the present disclosure.

[0062]

[0062] In addition to the above, the present specification further discloses a computer system including a processor, a memory, wherein the memory includes a computer program product according to the above or a non-transitory computer readable medium according to the above.

[0063]

[0063] In this disclosure, certain terms and technical specifications are used, some of which are explained with reference to the figures and some of which are explained and specified herein.

[0064] Cost Function or Cost

[0064] The term cost function or cost in relation to radiation therapy describes a measure of distance or difference between the obtained plan quality and the desired plan quality. The cost function may include the clinical objective of radiation therapy planning and represents a key point of the optimization process during planning. Various variables can be added to the cost or cost function, such as exposure of non-target tissue, non-target tissue exposed area (ENTA), bridge gap opening in the MLC and / or change of MLC or couch angle between segments or arc beams. Such variables may lead to additional costs. The value provided by the cost or cost function may be considered as a quality indicator of the treatment plan. Typically, the cost function for radiation therapy planning goes through an optimization procedure. The cost or cost function may be based on dose level criteria as a physical cost function or on biological criteria as a biological cost function. The objective of the cost function is generally to provide a treatment plan with a highly conformal dose distribution without irradiating non-target tissue exposed areas or organs at risk. In many solutions, the cost function or cost is optimized by finding the minimum of such a cost function. Costs, cost functions or ENTA costs can be very complex, but as the heart of the invention does not lie in determining the costs or cost functions, the explanation provided is sufficient to understand the aspects disclosed herein.

[0065]

[0065] The cost function may include the number of arc paths used to treat the target set. In general, the more arc paths used in the treatment, the higher the cost, since each arc path adds time to the treatment. Thus, if a certain cost level A can be achieved with seven arc paths and a first treatment plan, and the same cost level A can also be achieved with only six arc paths and a second treatment plan, the second treatment plan with six arc paths is more advantageous and therefore receives a lower total cost than A. The cost function that depends on the number of arc paths with a single output value may be designed in the following manner.

[0066]

[0066] Let T be a threshold, and all ENTA costs below T are considered equal. The cost is defined as follows:

[0067] c(P,N;T)=10 5 *round(c ENTA (P;T),3)+N

[0068]

[0068] where P is the target partition, N is the number of arc paths (up to 99), and round(x,3) means to round the decimal part following the integer part of x to three digits. That is, the function can be formulated as follows:

[0069]

[0069] An example cost is calculated as follows: First, calculate the value of the ENTA portion of the cost. If the resulting value is below a threshold T (T is equal to or greater than zero), set the value to T exactly. Round the value to a maximum of three decimal places, then multiply the result by 100,000. This ensures that you have an integer with two trailing zeros. Finally, add the number of arc paths in the treatment plan (less than 100) to get the total cost. Select the target segment with the lowest total cost as the optimal target segment.

[0070] Alternatively, a cost function that takes into account the number of arc paths can also be designed in the following manner: First, if the cost of a section is below a threshold T (T is equal to or greater than zero), set its cost exactly to T. After this step, select the target section with the smallest cost. If multiple target sections have the same cost (e.g., cost T), sort the target sections by the number of arc paths in the final treatment plan, and select the treatment plan with the smallest number of arc paths in the final treatment plan as the optimal target section.

[0071] Multileaf Collimator

[0071] A multi-leaf collimator (MLC) includes a frame with a rectangular aperture and many leaf pairs arranged adjacent to each other along opposite sides of the aperture. Two leaves of a leaf pair are arranged opposite each other and can be moved to completely close a portion of the aperture or to expose all or part of that portion of the aperture. Each leaf pair defines a linear portion of the MLC. There are various techniques for calculating the movement pattern of the MLC during beam delivery. For example, in sliding window delivery, the leaves move in one direction through the field, and the distance between the opposing leaves is selected to pass radiation in areas to be exposed to radiation and block other areas for a time determined by the fluence map. Multiple sliding window leaf sweeps can be performed in sequence without stopping delivery, resulting in a movement pattern in which the leaves move back and forth over the treatment area.

[0072]

[0072] Another example of radiation delivery is step-and-shoot delivery, where the radiation beam or leaves of the MLC are stationary while the beam is on.

[0073]

[0073] In the case of an arc beam, the leaves of the MLC can typically move in either direction while the beam is on.

[0074]

[0074] The MLC can be rotated to various angles around the beam central axis to limit the beam in the most suitable way depending on the patient geometry. A given rotation of the MLC with respect to the beam central axis is called the collimator angle. Because the patient geometry changes depending on the beam direction and / or gantry angle, it is also possible to rotate the MLC to different collimator angles at different gantry angles. The MLC can also be rotated during the delivery of a static beam, i.e., the collimator angle can be a function of the delivery time, gantry angle, or cumulative monitoring units (MU) of the beam. In the current common method, the collimator angle is manually selected and kept constant throughout the arc beam or static beam.

[0075] Target Area

[0075] In this specification, the target partition may refer to a division of targets into groups, regardless of the candidate arc paths. The target partition typically includes several target groups. In light of this, the target partition may differ with respect to the target set and the various target groups. Thus, the target set may be divided into various target groups, and thus into various target partitions. In addition, the target partition may also refer to a division of targets by candidate arc paths. The target partition typically includes target groups, whereby each target is part of at least one target group, but in some cases, particularly when the total cost or ENTA is lower, the targets may not be part of at least one target group each. Typically, at least one target partition is examined per candidate arc path. However, it may be multiple target partitions, which may be specified in the manner disclosed herein.

[0076] Candidate Arc Path

[0076] A candidate arc path in the context of the specification of this specification describes and means a path along which the MLC travels over the patient, which path is typically arc-shaped due to the configuration and / or design of the radiation therapy system. The candidate arc path in a way influences or determines the potential gantry angle, couch angle and / or even collimator angle. When treating a target group, it is important to select a candidate arc path relevant to the treatment in order to reduce the ENTA cost of the patient and provide an efficient plan of the radiation treatment. The aperture of the MLC and the associated cost of the ENTA are correlated with the selected candidate arc path, and by iteratively performing target grouping and target segmentation while correlating with the candidate arc path, both the outcome for the patient and the efficiency of the treatment are significantly improved. The selection of the target segmentation or grouping depends on the candidate arc path provided. The quality of a particular target section may depend on the candidate arc path. The candidate arc path may be provided via a treatment planning system or a physically active person, such as a medical professional. Typically, the candidate arc-paths for or considered for the methods disclosed herein are allowed to deviate to some extent from the provided candidate arc-path or paths. In some cases, the candidate arc-paths may correspond to couch angles. A current candidate arc-path herein further refers to a candidate arc-path currently being examined and / or calculated and / or analyzed.

[0077] Arc Beam

[0077] In this specification, an arc beam is considered to be a beam of radiation that can pass through the MLC at a specific location along the trajectory of the arc path. The arc beam can be a beam that occurs instantaneously at a fixed point along the trajectory, or it can be a continuous arc beam that moves along the trajectory while adapting the collimator leaf or collimator aperture according to the target. In some cases, the arc beam is a static beam that occurs at a specific fixed interval along the trajectory of the MLC and the arc path or candidate arc path. The arc beam is also shown in the figures of this specification.

[0078]

[0078] The method according to the present invention according to the first, second or third aspect can be integrated into any known treatment planning system to improve treatment planning and reduce the burden on the treatment planner. The method according to the first, second or third aspect can be used and applied in radiation treatment planning independently of each other, but they can also be used in combination with another.

[0079] BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In the following the invention will be explained in more detail, by way of example, with reference to the accompanying drawings, in which: [Brief description of the drawings]

[0080] [Figure 1]

[0079] A schematic diagram of a radiation treatment setup is shown. [Diagram 2]

[0079] A schematic diagram of a target and the arc path traveled by an MLC on a gantry for radiation treatment of the target is shown. [Figure 3a]

[0079] Schematic showing the principle of target division or grouping through a patient's head with seven targets. [Figure 3b]

[0079] Schematic showing the principle of target division or grouping through a patient's head with seven targets. [Figure 3c]

[0079] Schematic showing the principle of target division or grouping through a patient's head with seven targets. [Figure 3d]

[0079] Schematic showing the principle of target division or grouping through a patient's head with seven targets. [Figure 4a]

[0079] The principle of target clustering is illustrated diagrammatically. [Figure 4b]

[0079] The principle of target clustering is illustrated diagrammatically. [Figure 5a]

[0079] Schematically illustrates the principle of using target groups to avoid bridge gap opening in a multi-leaf collimator (MLC). [Figure 5b]

[0079] Schematically illustrates the principle of using target groups to avoid bridge gap opening in a multi-leaf collimator (MLC). [Figure 5c]

[0079] Schematically illustrates the principle of using target groups to avoid bridge gap opening in a multi-leaf collimator (MLC). [Figure 6a]

[0079] Schematic of the principle of avoiding non-target exposed areas (ENTA). [Figure 6b]

[0079] Schematic of the principle of avoiding non-target exposed areas (ENTA). [Figure 7a] Illustrated diagrammatically is the principle of jagged leaf aperture which the present invention seeks to avoid. [Figure 7b] Illustrated diagrammatically is the principle of jagged leaf aperture which the present invention seeks to avoid. [Figure 8]

[0079] Another embodiment according to the present invention is illustrated generally, this embodiment relating to a dose control shield. [Figure 9] 1 illustrates diagrammatically an embodiment of a method according to a second aspect of the present invention; [Figure 10] 10 illustrates, in schematic form, another embodiment of the method according to the first aspect of the present invention. [Figure 11] FIG. 13 shows another embodiment of a method similar to that of FIG. 10, which uses all candidate arc paths. [Figure 12] 10 illustrates, in schematic form, yet another embodiment of the method according to the third aspect of the present invention. [Figure 13]

[0079] A schematic diagram of a radiation therapy treatment system. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0081] Detailed Description

[0080] Turning now to diagrams illustrating different aspects and concepts of the present invention, FIG. 1 shows a schematic diagram of a radiation therapy treatment system 1 including a gantry 2 having a multi-leaf collimator (MLC) 4. The radiation therapy treatment system 1 may be used for stereotactic radiosurgery (SRS). The radiation therapy treatment planning system 1 may further include a patient couch or couch 6 used to position such a patient during the delivery of radiation therapy to such a patient. For illustrative purposes, the radiation therapy treatment system 1 is shown without a patient. The couch 6 may be rotated about an axis to position the patient in an optimal position during or for treatment. In the example shown in FIG. 1, the couch angle 8 may be adjusted by rotating the couch 6 about the z-axis. Other movements of the couch 6 may also be possible. The MLC 4 may be rotated about a collimator axis c to adjust the collimator angle 12. The gantry 2 may be rotated about a rotation center 13, so that the collimator axis c depends on the position of the gantry 2 and thus the gantry angle 10. From the MLC 4, a radiation beam or arc beam 15 can be extended towards a part of the body 17 containing a target or target volume 18. The radiation beam 15 travels along a collimator axis c, but its shape in cross section is perpendicular to the collimator axis c and depends on the MLC aperture as explained later in this specification (see Figs. 5a-7b). The gantry can rotate a full 360° around a center of rotation 13 to irradiate the target (not shown in Fig. 1) from different gantry angles 10. The gantry 2 can rotate around the y-axis as shown in Fig. 1. To better reach the target, the MLC 4 can also be mounted on an extension arm 14 or the like, which is connected to or part of the gantry 2.

[0082]

[0081] It should be noted that the angles can be defined in various ways. For illustrative purposes, in this specification, the couch angle 8 is measured counterclockwise from the position shown in Figure 1, which means that in the view shown in Figure 1, the couch angle 8 is 0°. The gantry angle 10 can be similarly measured counterclockwise from the horizontal plane defined by the x-axis and the y-axis, which means that the gantry angle 10 in Figure 1 is about 110°. The collimator angle 12 can be measured starting from a line defined by the longitudinal axis d of the extension arm 14.

[0083]

[0082] During its path along the various gantry angles 10, the MLC assumes different positions along its arc path. Such an arc path is shown and described in FIG.

[0084]

[0083] Figure 2 shows a configuration for treating a patient's head 16 with three metastases 18' or targets 18' by radiation therapy treatment. For illustrative purposes, only the patient's head 16 is shown. In Figure 2, several arc paths 20, 20', 20'' are shown. In addition, in Figure 2, another arc path 22 is also shown. Arc path 22 corresponds to a gantry angle of 0° to 360° and a couch angle of 0°. Arc path 20 corresponds to a gantry angle of 0° to 180° and a couch angle of about 45°. Arc path 20' corresponds to a gantry angle of 0° to 180° and a couch angle of about 90°. Finally, arc path 20'' corresponds to a gantry angle of 0° to 180° and a couch angle of about 120° or 240° (360°-120°) depending on the definition of the couch angle.

[0085]

[0084] To further explain the radiation procedure, the arc path 20 is used here, which corresponds to a gantry angle of 0°-180° and a couch angle of about 45°, as explained above. The arc path 20 is used to treat one of the targets 18' via the arc beam 24. For illustrative purposes, only one arc beam 24 is shown along the arc path 20. Additionally, and for illustrative purposes, only one of the targets 18' is irradiated via the MLC aperture 26a. The MLC 4 and, therefore, the MLC aperture 26a are positioned at a collimator angle 12 of about 15°-20°. The gantry angle 10 of the example shown in FIG. 2 of the arc beam 24 is about 30°. The arc beam can be delivered to the target 18' from various positions at intervals along the arc path 20. These intervals can be regular intervals along the arc path. In conformal radiation therapy, the arc beam 24 is delivered from at least some positions 28. Between two adjacent positions 28, the arc beam is typically delivered and the leaves 40 of the MLC 4 are not closed. However, in some radiation therapy methods, the arc beam is not delivered along at least a portion of the arc path 20, particularly during movement of the MLC 4, but the MLC aperture is not closed. The invention described herein is useful for a variety of solutions including continuous arc beam delivery along the arc path 20, particularly using conformal arc therapy, but also for VMAT, and more generally for intensity modulated radiation therapy (IMRT), including IMRT using static field or static beams.

[0086] Between two adjacent positions 28 or control points 28, the leaf 40 typically moves linearly between the specified positions / MLC openings. In some cases, the movement of the leaf 40 may be non-linear, for example due to acceleration of the MLC 4 during rotation.

[0087]

[0086] The locations 28 shown in FIG. 2 may also be referred to as segments and may instead be used as control points 28 during treatment simulation and planning. Each control point 28 may be used to determine the optimal collimator angle 12, the optimal MLC aperture 26a, and may also optimize the couch angle 8 by plus or minus a few degrees to optimize radiation delivery. In addition, the arc path 20 may also be optimized by a few degrees during calculation to find the best arc beam, arc path, and therefore arc plan for radiation delivery. The control points 28 may be spaced at regular intervals along the entire arc path 20. In the example of FIG. 2, the control points 28 are spaced at gantry angles 10 of 5°-7°. Any other intervals between 1°-25°, preferably between 2°-8°, are also possible and feasible. For illustrative purposes, only four control points 28 are shown in FIG. 2, but the reader will understand the concept of distributing the control points 28 along the arc path 20 during calculation and determination of the treatment plan. The control points 28 may also be used to determine whether any of the following concepts and aspects can be optimized:

[0088]

[0087] Figures 3a-3d show aspects of the present disclosure related to target divisions or target compartments. Figures 3a-3d show a patient's head 16', including an ear 30, used for illustrative purposes. The patient has a total of five metastases or targets 18a-18e on the head 16'. For illustrative purposes, in Figure 3a, only metastases or targets 18a-18e are shown as reference numbers. However, Figures 3b-3d disclose the same head 16' with the same metastases or targets 18a-18e, but these Figures 3b-3d illustrate the concept of target compartments 36, 36', 36'' and target groups 34, 34', 34''.

[0089]

[0088] Turning now to Fig. 3b, it can be seen that when grouping the five targets 18a-18e into target groups 34, 34', 34'', some conditions may apply. In this example, these conditions are that each target section 36 may contain at most three target groups 34, 34', 34'', whereby less than three target groups 34, 34', 34'' are allowed (see Fig. 3c). However, another condition, not shown in Figs. 3b, 3c, 3d, is the physical distance. Thereto, it may be specified that in one target group 34, 34', 34'', the targets 18a-18e located furthest from each other may only be spaced apart by a certain distance. This distance may be specified as a condition prior to the target grouping and target division steps. This distance may be between 0.5 cm and at most 8 cm, 9 cm or 10 cm. In FIG. 3b, a first target group 34 includes targets 18a and 18b, a second target group 34' includes target 18c, and a third target group 34'' includes targets 18d and 18e. Thus, only one target 18c can be positioned in target group 34'. From FIG. 3b, it can be seen that target compartment 36 can be selected in other ways as shown in FIGS. 3c and 3d. The target compartment or target compartments are thus a division of the target into target groups, each target compartment representing a different target group compartment or division. Target compartment 36' shown in FIG. 3c includes two target groups, and target compartment 36'' shown in FIG. 3d includes three target groups. Those skilled in the art will understand that there are many more possible target compartments than 36, 36', 36'', and many target compartments can be directly excluded because it is impractical or impossible to treat such target compartments (not shown). The methods disclosed herein consider all or at least some of the possible target compartments using predefined conditions such as the maximum number of allowable target groups 34, 34', 34" within a target compartment 36, 36', 36" and the maximum distance between two targets that are located furthest from each other but still within one group. Alternatively, this maximum distance can be specified or given as a maximum volume.Target fractionation further takes into account other parameters when optimizing multi-meta treatment, such as the concept of not exposing so-called non-target exposed areas (ENTA) to radiation, as explained with reference to figures 5a-6b. Figures 3a-3d refer to a first embodiment of the present invention.

[0090] 4a and 4b, another aspect of the present invention, called target clustering, is disclosed and explained. Here, a different head 16'' of a patient having seven metastases 18f-18l or targets 18f-18l is used to explain the concept or second aspect of the present disclosure.

[0091]

[0090] It should be noted that seven targets 18f-18l are a rather large number, but cases with up to 15 metastases have been treated. This means that multi-meta cases with up to 15 or even more targets can occur, and here we explain the concept of clustering based on the seven targets 18f-18l. The seven targets 18f-18l are difficult to treat in exactly the same arc, and therefore in the same isocenter. In the exemplary case shown in Figures 4a and 4b, the targets 18f-18l are too far apart to be safely treated in the same arc / isocenter due to uncertainties in rotation, which means that the targets 18f-18l need to be split into clusters 38, 38' so that each cluster 38, 38' can be treated separately, for example by using separate or independent isocenters / arcs. The clustering step thus dividing the targets 18f-18l into clusters 38, 38' may be used in the radiation therapy methods disclosed herein, such as conformal treatment or IMRT, or it may be used in other radiation therapy treatment methods, such as VMAT. The clusters 38, 38' may be selected based on the maximum volume allowed for the targets, or they may be selected based on geometrical constraints. Another further parameter that may be considered for selecting the clusters 38, 38' of targets may be the maximum number of targets for which target splitting and target grouping may be allowed. Such constraints may be selected at four, five or six targets. The clustering step or method disclosed in Figs. 4a and 4b may be used in conjunction with the present invention, or it may be used as an independent method in other radiation therapy planning methods or treatments to improve the results of radiation therapy planning.

[0092]

[0091] Figure 5a discloses an MLC aperture 26b including a leaf pair 40, which is designed to shield radiation coming from a radiation source (not shown in Figure 5a) in the MLC 4. In Figure 5a, five targets 18m-18q are shown, here in the form of spherical targets. In Figure 5a, irradiation of all targets 18m-18q is performed, not in practice but as a simulation. Analysis of the resulting irradiation area results in an ENTA 42, called a bridge gap aperture 42. This area between targets 18p and 18o is not irradiated during treatment, but would be irradiated if the MLC aperture 26b shown in Figure 5a is used. As shown in FIG. 5b, by grouping targets 18m-18q into a target section 36''' including two groups 34''', 34'''', a first group 34''' including targets 18n and 18o and a second group 34'''' including targets 18m, 18p and 18q, bridge gap openings 42 can be avoided, and thus ENTA 42 can be avoided. The method according to the present disclosure takes such ENTA 42 into account when finding the best possible target section 36'''', as shown in FIG. 5b. In FIG. 5b, a radiation treatment or at least a simulation of a treatment or MLC opening 26c of the second group 34'''' is shown, and in FIG. 5c, a radiation treatment or at least a simulation of a radiation treatment or MLC opening 26d of the first group 34'' including targets 18n, 18o is shown.

[0093] 5b and 5c show how leaf pairs 40 are closed over the non-treated target groups 34''', 34'''', while opening over the treated targets or target groups 34''', 34''''. It is further noted that FIGS. 5a-5c do not show the change in collimator angle 12 for ease of illustration. However, it is conceivable and within the scope of the present invention that the target splitting step may include checking and considering various collimator angles 12 at the control points 28 (see FIG. 2). In FIG. 5b, treatment of the second target group 34'''' is performed, and in FIG. 5c, treatment of the first target group 34''' is performed, thereby avoiding exposure or ENTA 42 (see FIG. 5a).

[0094]

[0093] Another example of avoiding the ENTA 42 region is shown in Figures 6a and 6b, which show another MLC aperture 26e, which may show a simulated or actual treatment showing a banana-shaped target 18r or tumor 18r. As shown in Figure 6a, the leaf pair 40 is open over the target 18r while the collimator angle is 0° (not shown). For illustrative purposes, this angle is assumed to be 0°, but it can be any other angle depending on how the collimator angle is defined. Figure 6a shows how radiation would be exposed (simulated) or exposed (actual treatment) to the ENTA 42' at the current collimator angle. As shown in Figure 6b, the ENTA 42' due to the leaf pair configuration and MLC configuration can be avoided by using another MLC aperture 26f by rotating the MLC by about 90°.

[0095]

[0094] In the MLC aperture 26f of Figure 6b, the same target 18r as shown in Figure 6a is treated, but the MLC is rotated by approximately 90°. As mentioned, this avoids exposing the ENTA 42' to radiation. Such issues are also taken into account when methods according to the present disclosure evaluate the MLC angle 12 and optimize the treatment plan.

[0096]

[0095] As shown in Figures 5a-6b, the exposure of ENTA 42, 42' can be expressed in the treatment plan optimization by adding the ENTA cost to the optimization for each exposure of ENTA when the target compartments 36, 36', 36'', 36''' and therefore the various target groups 34, 34', 34'', 34''' are analyzed to find the optimal target compartment 36, 36', 36'', 36'''.

[0097]

[0096] In Figures 7a and 7b, further costs that can be added to the optimization procedure are shown. Figure 7a shows an MLC aperture 26g used for a target 18s having a certain shape. The MLC aperture 26g shown in Figure 7a is a conformal MLC aperture 26g in that the leaf pairs 40 are open over the entire target 18s, while in Figure 7b, the leaf pairs 40 of the MLC aperture 26h are not open over the entire target 18s, which means that the MLC aperture 26h in Figure 7b is not conformal but is a so-called jagged aperture. Jagged MLC apertures 26h are avoided in the method according to the disclosure of this specification. If jagged MLC apertures 26h occur despite the optimization being performed, such jagged MLC apertures 26h can be weighted in the optimization procedure by adding a cost for jagged MLC apertures at one or several control points 28 (see Figure 2). The reason that jagged MLC apertures 26h should be avoided is that such jagged MLC apertures 26h adversely affect the robustness of the dose delivery: there can be uncertainty as to whether the patient is positioned exactly as expected during the dose calculation, and therefore there can be uncertainty as to whether the dose calculation is correct.

[0098]

[0097] A challenge with conformal arc planning is that in some cases, the dose levels for different targets by the same arc beam are highly correlated, and therefore it is difficult to simultaneously obtain the desired dose levels for multiple targets, i.e., multiple targets within one target group. Some solutions propose VMAT to optimize the dose levels, but VMAT results in jagged leaf positions that need to be avoided in conformal arc planning or conformal radiotherapy treatment. Figure 8 shows a similar view to Figure 5b, but with a different MLC aperture 26e. Here, a second target group 34'''' including targets 18m, 18p, and 18q is exposed to radiation, but only targets 18m and 18p are exposed, and not target 18q. Target 18q is shielded from radiation at this particular control point 28 or arc beam 24 (see Figure 2), because the optimization procedure showed that target 18q would receive a dose level that would exceed the desired dose level without shielding. This procedure is referred to as dose control shielding and is disclosed herein as a third aspect of the present invention. One skilled in the art will appreciate that other targets in the second group 34'''' can be shielded with other arc beams 24 along the arc path 20 to achieve a desired dose level at targets in the target group 34, 34', 34''', 34''''.

[0099]

[0098] The optimization procedure according to the third aspect may include a segment weight optimization, also called control point MU optimization, in order to calculate and optimize the dose levels as much as possible. The difference between the actual dose levels at the targets 18a-18q and the desired dose levels or dose level targets is then calculated. The segment weights or arc beam MU are then increased so that all the actual dose levels of the targets 18a-18q are equal to or greater than their dose targets. The target 18q or targets having a dose level higher than the dose level target are then shielded over at least a part of the arc path 20 and thus with a part of the arc beam 24 or segment, according to the concept shown in Fig. 8. This is done by analyzing the arc beam 24 along the arc path 20 and modifying the arc beam 24 so that the desired dose level is achieved at each target 18q by modifying the configuration of the MLC 4 at some of the control points 28 by closing the leaves 40 of the MLC 4 over the relevant target 18q. This procedure allows the actual delivered dose level for each target to be reduced to the desired dose level.

[0100]

[0099] The procedure for dose control shielding according to the third aspect of the present invention can be performed after clustering according to Figures 4a and 4b has been performed, and also after target segmentation according to Figures 3a and 3d has been performed, taking into account the avoidance of bridge gap opening 42 or ENTA 42'. However, dose control shielding according to Figure 8 can also be performed in other treatment planning methods and is not limited to be used only in the first and second aspects disclosed herein.

[0101]

[0100] Having now described various aspects and concepts of the present invention with reference to Figures 1-8, methods according to aspects of the present invention will now be described with reference to Figures 9-11.

[0102]

[0101] Fig. 9 shows a method for clustering of targets. The clustering method may be performed before target division. As input S01 to the method, the maximum allowable distance between two targets of the same cluster is given. This maximum allowable distance may be a threshold value. As yet another input, the number N of specific targets may be given. N may be an integer, such as 4, 5 or 6 targets. Preferably, the maximum allowable distance is the main criterion for the selection of clusters. In addition, input S01 may be given as the maximum number M of clusters 38, 38' of the clustering step or method. Another potential input is the maximum distance between two targets in one cluster, so that clusters can be determined and selected based on geometrical conditions. Alternatively, the maximum volume may be used as input for the method. Once the method has received these conditions, it determines the number of targets to be treated (S02), including the relative position between the targets, and then checks whether the number of targets exceeds the number N of specific targets (S03). If the number of targets exceeds a certain number N, the targets are divided into clusters based on geometric parameters such as the maximum distance between two targets of the same cluster 38, 38', the volume in which all targets of one cluster 38, 38' must fit (S04a). If the number of targets does not exceed a certain number N, the method according to Fig. 10 can be performed.

[0103]

[0102] The method according to Figure 9 will now be explained with reference also to Figures 4a and 4b.

[0104]

[0103] Figure 10 shows a target splitting optimization method when several targets 18, 18', 18a-18q need to be treated during one radiation treatment. It is a treatment planning method for generating a radiation therapy treatment plan in which a set of targets 18a-18q is to be treated, the method using a multi-leaf collimator (MLC) 4 for shaping an arc beam 24 or radiation beam 15, a gantry 2 including the MLC 4, which can rotate at least partially around the patient, a couch 6 for positioning the patient, the MLC 4 defining a collimator angle 12, the gantry 4 being movable around a gantry angle 10, and the couch 6 being movable around a couch angle 8. The target splitting optimization method includes: - providing (S05) the number of candidate arc paths with isocenters, the maximum number of arc beams for the treatment plan, and the maximum number of target groups 34, 34', 34'', 34'''', 34'''' allowed for the treatment depending on the maximum number of arc beams; - providing (S06) the shape and position of each target 18a-18q of the target set; - optionally, a step (S07) of calculating for the target set at least some possible target partitions 36, 36', 36'', each of which comprises a target group 34, 34', 34'', 34''', 34'''', whereby each target 18a-18q is part of at least one target group 34, 34', 34'', 34''', 34'''', 34'''', and, optionally, target partitions comprising target groups that are determined to be non-optimal, i.e. target groups that add a high cost to the current target partition, can be directly ignored from the possible target partitions; - calculating (S08) target partitions 36, 36', 36'' of at least some possible target partitions 36, 36', 36'' of at least some possible target partitions of the target set based on the current candidate arc paths 20, 20', 20'', 22, the at least some target partitions 36, 36', 36'' optionally including a maximum number of target groups 34, 34', 34'', 34''', 34'''', 34'''' according to a maximum number of target groups for all arc beams / number of arc beams constraint; - determining (S09) a cost for each target group 34, 34', 34'', 34''', 34'''' of the possible target section 36, 36', 36'', taking into account the current candidate arc paths, at least one gantry angle 10 and at least one MLC angle 12 for each target group 34, 34', 34'', 34''', 34'''' of the possible target section 36, 36', 36'', - determining (S10) the cost of the possible target partition 36, 36', 36'' and the current candidate arc path by summing the costs of the target groups 34, 34', 34'', 34''', 34'''' of the possible target partition 36, 36', 36'', - a step (S11) of determining whether at least some of the target sections 36, 36', 36'' have all been considered; - repeating the calculation, determination and summation steps for each possible target partition and each candidate arc path (S12); - selecting (S13) the optimal target segments 36, 36', 36'', arc paths 20 and arc beams 24 that have the smallest total cost of the treatment plan, taking into account the maximum number of target groups across all arc beams; Includes.

[0105]

[0104] The method shown in Fig. 10 and described in the preceding paragraphs relates to and makes use of the concepts of arc paths 20, 20', 20'' and control points 28 shown and described with reference to Fig. 2, the concept of target splitting shown and described with reference to Figs. 3a-3d, and the concept of avoiding such ENTAs 42, 42' by defining the cost of such ENTAs 42, 42' shown and described with reference to Figs. 5a-6b. In addition, although not explicitly described in the method according to Fig. 10, the method is designed to be used in conformal treatments, thereby avoiding jagged MLC apertures shown and described with reference to Figs. 7a and 7b.

[0106]

[0105] A constraint on the above method may be to provide a total number of target groups across all arc beams, which may facilitate considering all potential arc beams at once.

[0107]

[0106] Each arc beam 24 along the arc path 20 takes into account the collimator angle 12 along the arc path 20 and / or at each control point 28.

[0108]

[0107] In an embodiment, the calculation step (S08) may consider all possible target compartments.

[0109] From the method according to FIG. 10, an arc plan can be provided based on the optimal target section 36, 36', 36''. Such an arc plan can include a series of arc paths 20, 20', 20'' used to treat the target group 34, 34', 34'', 34''', 34'''', said arc paths including information on the collimator angle 12, couch angle 8 and MLC apertures 26a-26i as well as other parameters for each arc beam 24 or control point 28. The arc plan provided by the method according to FIG. 10 can include a minimum cost arc path or minimum arc beam based on the optimal target section 36, 36', 36''. The arc plan can further include a calculated or desired dose for each target.

[0110]

[0109] Figure 11 shows a method scheme similar to that of Figure 10 for a target splitting optimization method when several targets 18, 18', 18a-18q need to be treated during one radiation treatment. It is a treatment planning method for generating a radiation therapy treatment plan in which a set of targets 18a-18q is to be treated, using a multi-leaf collimator (MLC) 4 for shaping an arc beam 24 or radiation beam 15, a gantry 2 including the MLC 4, which can rotate at least partially around the patient, a couch 6 for positioning the patient, the MLC 4 defining a collimator angle 12, the gantry 4 being movable around a gantry angle 10, and the couch 6 being movable around a couch angle 8. The target splitting optimization method includes the following steps: - providing (S05) the number of candidate arc paths with isocenters, the maximum number of arc beams for the treatment plan, and the maximum number of target groups 34, 34', 34'', 34'''', 34'''' allowed for the treatment depending on the maximum number of arc beams; - providing (S06) the shape and position of each target 18a-18q of the target set; - optionally, a step (S07) of calculating for the target set at least some possible target partitions 36, 36', 36'', each of which comprises a target group 34, 34', 34'', 34''', 34'''', whereby each target 18a-18q is part of at least one target group 34, 34', 34'', 34''', 34'''', 34'''', and, optionally, target partitions comprising target groups that are determined to be non-optimal, i.e. target groups that add a high cost to the current target partition, can be directly ignored from the possible target partitions; - calculating (S08) target partitions 36, 36', 36'' of at least some possible target partitions 36, 36', 36'' of the target set based on the current candidate arc paths 20, 20', 20'', 22, the at least some possible target partitions 36, 36', 36'' including a maximum number of target groups 34, 34', 34'', 34''', 34'''', 34'''', taking into account candidate arc paths for each target group 34, 34', 34'', 34'''', 34'''', at least one gantry angle 10 and at least one MLC angle 12 of the possible target partitions 36, 36', 36''; - determining (S09) a cost for each target group 34, 34', 34'', 34''', 34'''' of the possible target section, taking into account the current candidate arc paths 20, 20', 20'', 22, at least one gantry angle 10 and at least one MLC angle 12 for each target group 34, 34', 34'', 34''', 34'''' of the possible target section 36, 36', 36'', and - determining (S10) the cost of the possible target partition 36, 36', 36'' and the current candidate arc path by summing the costs of the target groups 34, 34', 34'', 34''', 34'''' of the possible target partition 36, 36', 36'', - determining (S11) whether at least some of the possible target segments 36, 36', 36'' and all of the candidate arc paths have been considered; - providing candidate arc-paths, providing shapes and positions, and optionally repeating the steps of computing, calculating, determining and summing for each possible target partition and each or at least some of the candidate arc-paths (S14); - selecting (S13) the optimal target sections 36, 36', 36'', arc paths 20 and arc beams 24, which have a minimum total cost of the treatment plan, and optionally taking into account limitations or constraints related to the maximum number of all target groups when summed over at least some of the target sections; Includes.

[0111]

[0110] The above method allows each candidate arc path and at least some of the target segments to be considered at least approximately simultaneously for the optimization step, as shown in Figure 11, thereby aiming for the best possible treatment plan.

[0112]

[0111] Figure 12 further illustrates a method according to a third aspect of the invention relating to a dose control shielding method. The dose control shielding method comprises: - determining (S14) an arc plan based on a minimum cost arc beam 24 or a minimum cost arc path 20, 20', 20'', the arc plan including a calculated dose for the treatment; - providing a dose level target for each target of the target group 34, 34', 34'', 34''', 34'''' of the optimal target section 36, 36', 36'', and determining the exposure time and monitoring unit for each target such that each target of the optimal target section 36, 36', 36'' obtains at least the desired dose level (S15); - identifying targets whose dose levels exceed a dose level target (S16);

[0113]

[0112] Calculating the required dose control shielding (S17) for targets having dose levels above the dose level target in a particular arc beam 24 or part of the arc beam 24, where the leaves of the MLC aperture are fully closed; Includes.

[0114]

[0113] The method according to Fig. 11 may be used with the invention disclosed herein or it may be used in another treatment planning method. Advantageously, the dose control shielding method is used in conformal arc beam treatment or arc planning.

[0115] 13 further illustrates a radiation therapy treatment planning system 42 including a processor 44, a memory 46 connected to the processor 44, an optional additional memory 50 also connected to the processor 44 and / or the memory 46, and an input / output interface 48 having a display and / or keyboard for user interaction. The processor 44 may be any combination of a central processing unit (CPU), a multiprocessor, a digital signal processor (DSP), an ASIC (application specific integrated circuit) capable of executing software instructions stored on the memory and capable of performing the methods described according to the embodiments herein. The memory 46 and / or the additional memory 50 may include software code in the form of a computer program product 52 including computer software 54 or instructions including any method according to the present disclosure. The computer software 54 may be executed by the processor 44.

[0116]

[0115] The memory 46 and the additional memory 50 may be either a random access memory (RAM), a read only memory (ROM) or any combination thereof. EXAMPLES

[0117] Working Example Example 1

[0118]

[0117] Assume that four candidate arc paths are provided to treat a set of targets. In this example, it is assumed that each target is treated from all four couch angles corresponding to the provided candidate arc paths. Furthermore, in this Example 1, it is assumed that each target is treated exactly once from each such couch angle / candidate arc path.

[0119] The objective is to find the best possible plan using a maximum of 10 arcs in the final plan, which means that the total number of target groups is a maximum of 10 across all candidate arc paths. 2While the number of arcs may be reduced to e.g. 9 or 8, such optimization must also be taken into account, if, according to e.g. clinic guidelines, it is acceptable to use 10 arcs, but it would be preferable to use fewer arcs if the dose of radiation to non-target tissues were not significantly increased.

[0120]

[0119] The optimization of this problem must be done simultaneously for all candidate arc paths. The reason is that if we try to optimize one arc at a time, we may find an optimal division into, for example, two, three or four target groups for the first candidate arc path. Since increasing the number of target groups reduces ENTA (if still not satisfactory), it may be desirable to divide the target set into four target groups for the first candidate arc path. Meanwhile, only six target groups remain for the remaining three candidate arc paths. If we divide into only two target groups, eight target groups remain, but with a higher ENTA contribution. However, at this point, we do not know the best configuration of candidate arc paths and target divisions, which means we cannot know what the optimal choice is.

[0121]

[0120] On the other hand, for simultaneous optimization, we start by computing the optimal partition for each of the four candidate arc-paths into, say, two, three, four target groups, thereby saving the cost of the optimal partition for each candidate arc-path and number of target groups. From here, we loop through all ways in which four numbers (one for each candidate arc-path) that are all either 2, 3, or 4 can be combined to sum to 10. Some examples are 2+2+2+4, 2+3+2+3, 3+3+2+2, etc. These options are then evaluated by summing the ENTA contributions for each candidate arc-path. In the case of 2+2+2+4, we sum the ENTA of the optimal way to split each of the first three candidate arc-paths into two target groups, and the ENTA of the optimal way to split the fourth part into four target groups. By doing this for all possible combinations and choosing the smallest sum, we obtain the optimal way to optimally split the targets for all candidate arc-paths. This optimal ENTA sum is then compared to a threshold. If it is low enough, i.e. ENTA is below the threshold, we can expect that a total of 9 beams will still be low enough. To find that out, we repeat the same process as above, but limit the total number of target groups to 9 instead of 10. If the total ENTA is low enough for 9, we can do it again with 8, and so on. In this way, we end up with an optimal treatment plan that uses up to 10 candidate arc paths without significantly affecting ENTA, reducing the number of final arc paths when possible.

[0122] Example 2

[0123]

[0122] Again, assume there are four candidate arc paths, but this time assume that each target does not need to be treated by each candidate arc path, but rather by at least three candidate arc paths. However, the set of arc paths that treat one target A does not need to be the same as the set of arc paths that treat another target B. Note that this example works and is useful in other situations as well, such as when there are up to five candidate arc paths and one wants to treat each target from at least four or at least three candidate arc paths.

[0124]

[0123] Suppose we want to treat six targets with as few arc beams as possible while keeping the sum of ENTA below a threshold. The fact that one target can be treated from only three of the four available candidate arc paths opens up the possibility of using fewer beams while achieving a better or at worst equally good plan compared to manually selecting three of the four candidate arc paths and treating all targets using only those candidate arc paths.

[0125] To achieve this, we proceed as in the previous example to compute the cost of optimal partitioning of each of the candidate arc paths into 2, 3, 4 parts (1 or 5 parts are also possible), but this time we add another stage of optimization to select which of at least 3 of the 4 candidate arc paths will treat each target.

[0126]

[0125] In the previous example 1, the calculation of the cost of the optimal way to split into 2, 3, 4 target groups is all based on the assumption that all targets are included in one of the target groups of the partition. Here, these calculations are not only performed for the entire target set, but also for each subset of targets. In other words, for each subset of targets, find the optimal way to split the subset into 2, 3, 4 target groups, given the restriction that the subset contains enough targets.

[0127]

[0126] Assume there are six targets A, B, C, D, E, F. For each candidate arc-path, do the following:

[0128] Enumerate all subsets of the set S = {A, B, C, D, E, F} of all targets. One such subset is S1 = {A, B, D, F}, where C and E are missing. Another subset is S2 = {B, C}, and yet another subset is S3 = {A, B, C, D, E, F}, which is equal to S itself.

[0129]

[0128] For each subset, find the optimal partition into 2, 3, 4 target groups. For S1,

[0130]

[0129] ○ Two target groups: {{A,D},{B,F}}, total cost 2.1

[0131]

[0130] ○Three target groups: {{A,D},{B},{F}}, total cost 0.3

[0132]

[0131] Four target groups: {{A},{D},{B},{F}}, total cost 0.0

[0133]

[0132] For S2, which has only two targets, there is no point in splitting into three or four target groups, so we simply find the optimal way to split into two target groups.

[0134] After doing as before for each candidate arc-path, the results are compiled into an optimal treatment plan. This is done by solving the problem for many options for the total number of arc-paths, as in Example 1 above, where these options were 10 arc-paths, 9 arc-paths, and 8 arc-paths. The number of arc-paths is then minimized under the constraint that the total cost must be below a threshold.

[0135]

[0134] However, in this setting, the problem is more complicated. For this Example 1 case, suppose we are only considering the case that the total number of arcs in the final plan should be 10. Then, for each possible way of selecting the partition size for each candidate arc-path, such as 2+2+2+4, which sums to 10, we need to consider all the ways of selecting different subsets to consider in each partition. That is, we need to compute one cost for each combination of ways that subsets for different candidate arc-paths can be selected. For example, we need to compute one sum of costs if the selected subsets are S3, S3, S3, S3, one sum of costs if they are S2, S3, S3, S3, one sum of costs if they are S1, S3, S3, S3, etc. In doing this, we consider only combinations that meet the requirement of at least three initial arc-paths for one target. Therefore, since target E is not included in either S1 or S2, S1, S2, S3, and S4 are not considered, and for this reason, it is treated by only two candidate arc paths (i.e., the third and fourth arc paths that treat S3).

[0136] After doing this for all valid choices of subset combinations, we assign the lowest cost among the valid subset choices as the cost to split into 2+2+2+4 target groups. After doing this for all possible partition sizes for each candidate arc path totaling 10 arc paths, we find the best treatment method using 10 arc paths, treating each target with at least 3 candidate arc paths. As in Example 1 above, this can be repeated for different numbers of total arc paths to achieve the desired results.

[0137]

[0136] For example, the result in this case can be said to be to split the subsets S3, S1, S3, S4 (where S4 = {B, C, E}) into 3, 2, 2, 2 target groups, respectively, giving a total of 9 arc paths, leading to treatment with a cost of 0. Doing this, as in Example 1, with the condition that each target is treated with all candidate arc paths, is equivalent to treating with subsets S3, S3, S3, S3. Replacing S1 and S4 with instances of S3 increases the treatment costs of these target partitions when splitting each of them into 2 target groups. Thus, in this case, to achieve a cost of 0, the number of target groups of these partitions increases, and therefore the total number of arc paths in the plan also increases to 10 or 11 to achieve a cost of 0.

[0138]

[0137] The present invention has been described above with reference to the embodiments disclosed above and in the accompanying drawings. However, those skilled in the art will appreciate that embodiments other than those disclosed herein are contemplated and are also possible within the scope of the present invention as defined by the appended claims.

Claims

1. 1. A treatment planning method for generating a radiation therapy treatment plan in which a set of targets (18a-18q) are to be treated, using a multi-leaf collimator (MLC) (4) for shaping an arc beam (24), a gantry (2) for holding the MLC, the gantry (2) being at least partially rotatable around a patient, and a couch (6) for positioning the patient, the MLC (4) being movable and defining a collimator angle (12), the gantry (2) being movable and defining a gantry angle (10), and the couch (6) being optionally movable, thereby defining a couch angle (8), the method comprising: - providing (S05) a number of candidate arc paths (20, 20', 20'', 22) with isocenters, a maximum number of arc beams (24) for said treatment plan, and a maximum number of target groups (34, 34', 34'', 34''', 34'''') depending on said maximum number of arc beams; - providing (S06) the shape and position of each target (18a to 18q) of said target set; - calculating (S08) target partitions (36, 36', 36'') of at least some possible target partitions of said target set based on current candidate arc paths (20, 20', 20'', 22), said at least some possible target partitions (36, 36', 36'') comprising a maximum number of target groups (34, 34', 34'', 34''', 34''''), each of said at least some possible target partitions (36, 36', 36'') comprising a target group (34, 34', 34'', 34''', 34'''', and determining (S09) a cost for each target group (34, 34', 34'', 34''', 34'''') of said possible target section (36, 36', 36'') taking into account said current candidate arc paths (20, 20', 20'', 22), at least one gantry angle (10) and at least one MLC angle (12) for each target group (34, 34', 34'', 34''', 34'''') of said possible target section (36, 36', 36''); - determining (S10) the cost of said possible target partitions (36, 36', 36'') and of the current candidate arc paths (20, 20', 20'', 22) by summing up the costs of said target groups (34, 34', 34'', 34''', 34'''') of said possible target partitions (36, 36', 36''); - repeating (S12) said steps of calculating (S08), determining (S09) and summing (S10) for each of said at least some of the possible target partitions (36, 36', 36'') and at least some of said candidate arc-paths (20, 20', 20'', 22); - selecting (S13) the optimal target segment (36, 36', 36'') and candidate arc paths (20, 20', 20'', 22) for which the total cost of said treatment plan is minimal; A treatment planning method comprising:

2. The treatment planning method of claim 1 , wherein the calculation step (S08) further takes into account at least one couch angle (8).

3. 3. The treatment planning method of claim 1 or 2, wherein the maximum number of target groups (34, 34', 34'', 34''', 34'''') is provided by a treatment planner or is any predefined number between 2 and 30, preferably between 8 and 20 for all arc beams (24), alternatively between 1 and 6 target groups per arc beam.

4. 4. The treatment planning method according to claim 1, wherein the calculating step (S08) further comprises dividing the candidate arc-paths (20, 20', 20'', 22) into control points (28) spaced at regular intervals, each interval corresponding to a gantry angle segment, and at each control point (28) an optimal MLC angle (12) and / or an optimal couch angle (8) is calculated.

5. 4. The treatment planning method according to claim 1, wherein for each control point (28), an optimal MLC aperture (26a-26i) is calculated, whereby the optimal MLC aperture is selected based on a minimum cost and an optimal MLC angle (12) for the previous and next MLC angles (12) at the previous and next control points (28).

6. 4. The treatment planning method of claim 1, wherein the calculating step (S08) further comprises dividing the candidate arc-paths (20, 20', 20'', 22) into control points (28) spaced at regular intervals, each interval corresponding to a gantry angle segment, the MLC angles (12) being maintained at fixed positions for each control point (28), and a cost of each candidate MLC angle (12) being calculated for each control point (28) to find an optimal fixed MLC angle (12) for each candidate arc-path (20, 20', 20'', 22).

7. The treatment planning method of claim 4, wherein further cost is added if the MLC angle (12) needs to be changed between neighboring control points (28).

8. The treatment planning method of claim 2, wherein further costs are added if the couch angle (8) needs to be changed between neighboring control points (28).

9. The treatment planning method according to any one of claims 1 to 8, wherein further costs are added if the total number of target groups (34, 34', 34'', 34''', 34'''') in the current target section (36, 36', 36'') exceeds a certain threshold, such as 15 to 20 target groups.

10. The treatment planning method of any one of claims 1 to 9, wherein the maximum number of arc beams (24) corresponds to the maximum number of target groups (34, 34', 34'', 34''', 34'''').

11. The treatment planning method of any one of claims 1 to 10, wherein the arc beam (24) is a conformal arc beam.

12. 12. The treatment planning method of claim 1, further comprising the steps of obtaining a minimum-cost arc beam (") having a minimum-cost pair of MLC angles (12) and gantry angles (10) for each target group (34, 34', 34", 34'", 34"") of the optimal target section (36, 36', 36") and delivering the minimum-cost arc beam (24), whereby each arc beam (24) is matched to one target group (34, 34', 34", 34'", 34"") of the optimal target section (36, 36', 36") for treatment such that each target group (34, 34', 34", 34'", 34"", 34"") receives one arc beam (24).

13. A treatment planning method according to any one of claims 1 to 12, further comprising the step of using the treatment planning method according to any one of claims 1 to 12 as a starting point for intensity modulated arc therapy optimization.

14. The treatment planning method of any one of claims 1 to 13, wherein at least some of the possible target segments (36, 36', 36'') are considered for various candidate arc paths (20, 20', 20'', 22), and the treatment planning method provides an arc plan including an arc beam (24) and an arc path (20, 20', 20'', 22) based on the optimal target segment (36, 36', 36'') and a minimum cost arc beam (24), taking into account a restriction on the maximum number of arc beams (24), e.g. a maximum total number of target groups (34, 34', 34'', 34''', 34'''') for the maximum number of arc beams (24).

15. receiving (S14) an arc plan and a calculated dose for each target (18a-18q) in said optimal target section (36, 36', 36''); - providing a dose level target for each target (18a-18q) and adjusting said arc plan (S15) so that each target achieves at least said dose level target; - identifying (S16) said targets (18a-18q) that achieve a dose level that exceeds said dose level goal; - calculating (S17) for all targets of said optimal target section (36, 36', 36'') the required dose control shielding, in which at least some leaves (40) of said MLC (4) are completely closed for at least some of the arc beams (24) for the targets (18a-18q) whose dose levels exceed said dose level target, in order to optimize said dose delivery and thus said dose level target; The treatment planning method of any one of claims 1 to 14, further comprising:

16. 12. The treatment planning method of claim 11, wherein the control points (28) and / or the starting gantry angle (10) and the ending gantry angle (10) are modified to avoid any completely closed MLC apertures (26a-26i) at the control points (28) or segments where all targets (18a-18q) are occluded.

17. 17. The treatment planning method according to any one of claims 1 to 16, wherein a clustering step (S04a) is performed before the steps of claim 1 if the distance (S01) between two targets (18a-18q) exceeds a certain threshold or if the number of targets exceeds N targets (S03), N being an integer, and the clustering step comprises the steps of: selecting a maximum number M of clusters (38, 38') based on the number of targets (18a-18q), where M is an integer; limiting the number of targets in a cluster (38, 38') to less than or equal to N; and selecting the targets in the cluster based on a maximum allowed distance between two targets of the same cluster.

18. 18. The treatment planning method of claim 17, wherein as input the position and three-dimensional shape of each target (18a-18q) of the target set relative to other targets of the target set is provided and an operator specifies the integers N and M.

19. A computer program product (52) comprising computer readable means which, when executed on a computer, causes said computer to perform the method of any one of claims 1 to 18.

20. 21. A radiation therapy planning system comprising a processor (44), a memory (46), the memory (46) comprising the computer program product of claim 20, the radiation therapy planning system being designed to perform the method of any one of claims 1 to 18.

Images