Computer-implemented method for radiation therapy treatment planning, computer program product and computer system for performing the method

Abstract

A computer-based method for optimizing a patient's radiation therapy treatment plan is presented, wherein a treatment plan to be provided by one radiation set is optimized taking into account previously delivered doses delivered by different radiation sets, one of the radiation sets being external beam radiation therapy and the other being brachytherapy.

Description

Technical Field

[0001] The present invention relates to a computer-implemented method for radiotherapy treatment planning, and to a computer program product and apparatus for performing such a method. In particular, the present invention relates to treatment planning for radiotherapy involving both external beam radiation therapy and brachytherapy for the same patient. Background Art

[0002] Most radiation therapy treatments are provided as external radiation, which is delivered to the patient from an external source, called external beam radiation therapy or EBRT. This is typically delivered in multiple fractions, such as 15 fractions, 30 fractions, or more. Alternatively, radiation can be delivered from a source placed inside the patient's body. This is called brachytherapy and involves placing one or more needles or other types of instruments within a target area within the patient's body to expose the target area to radiation from the inside. This is typically performed in fewer fractions, such as 1 fraction or 3 fractions.

[0003] For EBRT treatments, planning is currently performed primarily as an inverse planning process in an optimization treatment planner. The goal is typically to achieve a minimum dose or a uniform dose across the entire target volume. Brachytherapy or BT planning requires the placement of instruments for delivering radiation within the target volume, either physically or in a virtual environment, and the dose is typically developed in a pre-planning process in which the position of the instruments and the dose resulting from the instruments are determined. In low-dose-rate BT, one or more sources (called seeds) are implanted into the target volume and typically remain there for the foreseeable future. On the other hand, in high-dose-rate BT and pulsed-dose-rate BT, radiation is delivered by a radioactive source that moves through a hollow channel in an implanted instrument, which consists of, for example, a needle, catheter, and applicator. Typically, the BT dose distribution is not uniform, but is concentrated around the implanted instrument.

[0004] Combining EBRT with brachytherapy in the same patient is known in the art. Typically, EBRT fractions are delivered first, followed by brachytherapy, but the reverse order is also possible. During the first treatment, the patient's geometry will generally change. In addition, the instruments inserted into the patient to deliver brachytherapy will change the shape of the target volume and surrounding tissue. For these two reasons, the two modalities will be delivered to different patient geometries. In cases where both brachytherapy and external beam radiation therapy are used on the same patient, two separate treatment plans are traditionally developed: one for the EBRT portion of the treatment and one for the brachytherapy portion of the treatment. These are typically based on a total dose and a predetermined dose split between the different types of treatment. For example, a total dose of 80 Gy may be set, with the EBRT treatment contributing 60 Gy and the brachytherapy contributing the remaining 20 Gy.

[0005] It is an object of the present invention to provide an improved planning method for radiotherapy treatments involving both EBRT and brachytherapy. Summary of the Invention

[0006] The present invention relates to a computer-based method for optimizing a patient's radiotherapy treatment plan, comprising the following steps:

[0007] a) determining a dose previously delivered to the patient by means of a first set of radiation based on at least a first medical image of the patient,

[0008] b) obtaining a first optimization problem comprising an optimization function designed to optimize a dose distribution resulting from the second set of radiations according to a dose criterion set for a total dose and a previously delivered dose, the total dose being the dose resulting from the second set of radiations and the previously delivered dose,

[0009] c) optimizing a treatment plan for delivering the radiation therapy by the second set of radiation by means of a first optimization problem, wherein the optimization function is an objective function or a constraint, and one of the first set of radiation and the second set of radiation is brachytherapy and the other is external beam radiation therapy.

[0010] This will improve the planning of the treatment to be delivered by the second set of radiation by taking into account the treatment already delivered by the first set of radiation.The previously delivered dose may be estimated or calculated by means of any suitable method.

[0011] In a preferred embodiment, the first optimization problem comprises an optimization function designed to optimize the dose distribution resulting from the second radiation set to match the difference between the desired total dose distribution expressed by the dose standard and the previously delivered dose. This will provide an actual total dose resulting from the combined dose delivered by the two radiation sets that is a good match to the desired total dose distribution.

[0012] The dose criteria are preferably based at least on earlier medical images acquired prior to the delivery of the first set of radiation.

[0013] According to a preferred embodiment, before step a), the method further comprises the following steps:

[0014] d) obtaining an initial optimization problem comprising an optimization function designed to optimize a total dose distribution as a combination of a first dose distribution to be provided by the first set of radiation and a second dose distribution to be provided by the second set of radiation;

[0015] e) Optimizing the treatment plan as a combination of the first radiation set and the second radiation set by means of the initial optimization problem.

[0016] This will further improve the planning by already taking both radiation sets into account when planning the dose to be delivered by the first radiation set.

[0017] The at least one first medical image may comprise at least one image of the patient acquired after delivery of the first portion. This will provide the most correct information about the actual patient geometry.

[0018] Alternatively or additionally, the at least one first medical image may include at least one simulated image based on an estimate of the patient's geometry after delivery of the fraction. This may be appropriate if, for some reason, it is not feasible to acquire new images of the patient after delivery of the fraction or upon insertion of the brachytherapy device.

[0019] In a preferred embodiment, the method comprises deforming at least one of the doses delivered by the first radiation set and the second radiation set, respectively, to obtain a common geometry of the treatment portion and integrating them using a biological model, the optimization function being a combination of a penalty for the accumulated dose and a penalty for a radiation set-specific dose from the second radiation set. Typically, the dose of the second radiation set will be deformed to the first medical image.

[0020] Preferably, robust planning is used to account for uncertainties in the placement of brachytherapy devices, EBRT delivery and / or determined delivered dose.How to achieve robust planning is known in the art.

[0021] The present invention also relates to a computer program product comprising computer readable code means, which, when run in a computer, is arranged to cause the computer to perform the method according to any one of the above embodiments. The computer program product may be stored on any suitable type of non-transitory storage medium.

[0022] The invention also relates to a computer system comprising a processor and at least one program memory, characterized in that the program memory stores a computer program as defined above.

[0023] In a preferred embodiment, the present invention relates to planning brachytherapy plans, wherein the cumulative dose already delivered to the patient by some other modality is also taken into account. This means that deviations from the planned dose to the target volume can be compensated, and excessive doses to organs at risk can be compensated by changing the brachytherapy treatment plan accordingly. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The present invention will now be described in more detail by way of example and with reference to the accompanying drawings, in which:

[0025] Figure 1a 、 Figure 1b and Figure 1c is a cross-sectional view of a medical image of a patient.

[0026] Figure 2 is a flow chart of a general embodiment of a method.

[0027] Figure 3 is a flow chart of a more specific embodiment of a method including EBRT followed by BT.

[0028] Figure 4 is a flow chart of a second more specific embodiment of a method including BT followed by EBRT.

[0029] Figure 5 is a schematic overview diagram of a computer system in which embodiments of the present invention may be implemented. DETAILED DESCRIPTION

[0030] External beam radiation therapy (EBRT) involves providing radiation to a patient in the form of a beam delivered from the outside. The radiation can be any type of radiation, including photons, electrons, protons, or other ions. Brachytherapy (BT) involves inserting some type of instrument into a target area within the patient's body and using the instrument to deliver radiation from one or more points within the target area. The instrument can include many small needles and / or catheters, one or more larger applicators, one or more seeds, or any combination of different types of instruments. Depending on the number and size of the instruments, the geometry of the target area and the surrounding patient will be distorted.

[0031] Due to the different nature of the two radiation sets, EBRT and BT are planned using different treatment parameters. Treatment parameters for EBRT treatment include beam and beam limiting device configuration. Treatment parameters for BT treatment include variables such as instrument position and dwell time. Each radiation set typically involves radiation delivered in one or more fractions, typically, but not necessarily, the number of EBRT fractions is higher than the number of BT fractions, and BT can even be delivered in a single fraction.

[0032] As discussed above, inverse planning using an optimization problem is common for EBRT planning but not traditionally used for brachytherapy. The dose distribution for EBRT can be expressed as

[0033] d EBRT =d EBRT (x EBRT ) (1)

[0034] And the dose distribution of brachytherapy can be expressed as

[0035] d BT =d BT (x BT ) (2)

[0036] where x EBRT , x BT is the treatment parameter for the corresponding treatment modality.

[0037] The present invention involves optimizing treatment parameters for both EBRT and BT treatments simultaneously. This means that the optimization problem can be expressed as Equation (3)

[0038]

[0039] where x EBRT are the treatment parameters used for the EBRT sub-part of the treatment, x BT is the treatment parameter for the BT subcomponent of the treatment. EBRT and d BT are the doses for the EBRT sub-fraction and the BT sub-fraction, respectively.Instead of the dose d, some other parameters related to the respective sub-fractions can be used.

[0040] Typically, the optimization consists of deforming the doses to a common geometry and accumulating them using a biological model, with the objective function being a combination of a penalty on the accumulated dose and a penalty on the radiation set-specific dose.

[0041] Figure 1a 、 Figure 1b and Figure 1c are simplified examples of medical images acquired at different points in the process of the present invention, as will be combined with Figure 2 discussed in more detail. Figure 1a 1 is a cross section 11 of a schematic medical image through the abdomen of a patient, wherein a target volume 13 and organs at risk 15 are indicated for treatment planning according to an embodiment of the present invention. Figure 1b The changes that such treatment may cause to the patient's geometry are schematically illustrated by a corresponding cross-section 11' of a medical image of the same patient after a first type of treatment.As will be appreciated, the target volume 13 has shrunk as a result of the treatment, which is generally a desired result. Figure 1c Here is a corresponding section 11 " of a medical image of the same patient with a needle inserted into a target volume to provide brachytherapy to the patient. The needle is shown as a small dot 17 within the target volume. As can be seen, this also changes the geometry of the target volume 13" and of the area of the patient surrounding the target volume.

[0042] Figure 2 is a flow chart of an overall method according to an embodiment of the present invention.

[0043] In a first step S21, an image of the relevant part of the patient is obtained, such as a combination of Figures 1a to 1c In step S22, a dose standard is determined for a total dose distribution to be delivered as a combined plan including both EBRT and brachytherapy.

[0044] In step S23, an optimization problem is defined based on the image or images and a dose criterion for the desired total dose. The dose criteria are set as is common in the art. They typically include a minimum dose for all voxels of the target volume and often a maximum dose for one or more organs at risk. For example, the dose criterion may specify a total dose of at least 60 Gy in each target volume voxel and that no more than 30% of the organs at risk receive a total dose of more than 40 Gy. The dose criterion may also include a partial or full dose distribution. In S24, the treatment plan is optimized using an optimization problem. The optimization problem includes an objective function, such as function (3) above.

[0045] In step S25, a portion of the treatment plan is delivered to the patient, and in step S26, a cumulative dose delivered to the patient from that portion of the treatment plan is estimated. The cumulative dose may be determined in any suitable manner. Methods for doing so are well known in the art and are typically based on at least one medical image, such as a plurality of fractionated images acquired during the delivery of the first portion of the treatment plan.

[0046] In step S27, a new image of the same portion of the patient is obtained to view the new patient geometry after the portion was delivered in step S25. If applicable, other modifications may be made, such as the insertion of a brachytherapy device, and an image reflecting the resulting geometry may be acquired. The new image may be an image of the patient acquired at this stage, or a composite image based on an estimate of the new patient geometry.

[0047] In step S28, the remaining portion of the treatment plan is optimized again using an inverse planning method based on an optimization function that will be discussed in more detail below. The plan takes into account the cumulative dose delivered by the previously performed portions. The optimization problem should then include an objective function according to the following formula:

[0048]

[0049] in is the delivered dose from the first set of radiation determined in step S26, and d RS2 is the dose to be delivered by the second set of radiation.

[0050] where g is another objective function, which may or may not be equal to f, and the delivered dose over all fractions (measured or estimated) from the first radiation set is used as a fixed background dose for the planning of the second radiation set.

[0051] In step S28, the remaining re-optimized portion of the plan is delivered to the patient.

[0052] Figure 3 FIG3 is a flow chart of an embodiment of a method of the present invention, wherein the overall treatment plan includes first performing EBRT treatment, followed by BT treatment. Input data S31 for the plan includes a current medical image of the patient, a dose standard for a desired dose distribution, and a prediction model of the patient's geometry after EBRT treatment. The prediction model can be a modified or synthesized medical image. The input data can also include a prediction model of the patient's geometry after EBRT treatment with an included EB device. The medical image can be a CT image or any other suitable image modality, such as an MRI or ultrasound image.

[0053] In method step S32, an optimization problem is obtained based on the dose criteria and the input data. The optimization problem includes an objective function based on the total dose of both the EBRT sub-portion and the BT sub-portion of the treatment and optionally the dose to be delivered for each radiation set according to equation (3). This typically involves making the dose d EBRT and d BTAt least one of the images is deformed to a common geometry and accumulated using an appropriate biological model. This can include a combination of penalties for cumulative dose and penalties for specific doses. Models for establishing a common geometry are known and typically involve deformable registration of images. Models for determining cumulative dose are also known to those skilled in the art. For example, the biological concept EQD2 can be applied to provide an estimate of the total effective dose.

[0054] In a subsequent method step S33, an optimization is performed based on the common geometry and the cumulative dose. The optimization problem comprises an objective function according to equation (3) above. As will be appreciated, the objective function can also be extended to depend on treatment parameters, and the optimization problem can also include constraints that depend on dose or treatment parameters. The output S34 from the optimization step S33 is a total treatment plan comprising a sub-portion for each radiation set (i.e., an EBRT sub-portion and a BT sub-portion). Each sub-portion comprises a portion of the dose to be delivered by the corresponding radiation set and the number of fractions in which it is delivered.

[0055] The EBRT sub-portion of the treatment plan is then delivered to the patient in step S35 and the actual delivered dose from that delivery is determined or estimated in step S36. Preferably, the situation after EBRT delivery is evaluated and used to improve the BT sub-portion of the treatment plan as outlined in the following steps.

[0056] In step S37, updated images of the patient after EBRT treatment are obtained. These include new images of the patient to account for geometric changes that occurred during EBRT treatment. They also include images of the patient with the BT device inserted, as the BT device will cause some amount of deformation of the target volume and surrounding patient geometry, depending on the type of device.

[0057] In a subsequent step S38, the BT sub-portion of the treatment is re-optimized taking into account the delivered dose from step S35 and the new image obtained in step S37. The optimization problem in this case comprises an objective function expressed as the following equation (2):

[0058]

[0059] in is the delivered dose determined in step S35.

[0060] The output from step S38 is a new optimized BT treatment plan S39, which is preferably delivered to the patient.

[0061] If targeted Figure 2As discussed, steps S36 to S39 may be performed without first performing the preceding steps, that is, BT planning may proceed even without an initial combined planning of the two prior to EBRT treatment.

[0062] Figure 4 FIG4 is a flow chart of a method in which the first set of radiation to be delivered is BT. Input data S41 for the plan includes a current medical image of the patient, a medical image of the patient with the BT device inserted, and a dose standard for the desired dose distribution. Preferably, the input data also includes a predicted model of the patient's geometry after the BT treatment. The medical image can be a CT image or any other suitable image modality, such as an MR or ultrasound image.

[0063] In method step S42, the optimization problem is defined in a similar manner to step S32. When first delivering the BT dose portion, images of the patient with the BT device inserted are already available as input. These images can be deformably registered to provide a geometric correspondence between the treatment geometries. As in step S32, the dose d EBRT and d BT The images are deformed to a common geometry and accumulated using an appropriate biological model. This can include a combination of penalties for accumulated dose and penalties for specific doses. Models for establishing a common geometry are known and typically involve deformable registration of images. Models for determining accumulated dose are also known to those skilled in the art.

[0064] In a subsequent method step S43, an optimization is performed based on the common geometry and the cumulative dose. The optimization problem includes an objective function according to the above equation (3). As will be appreciated, the optimization problem can also be extended to depend on treatment parameters and also include other objective functions and / or constraints. The output S44 from the optimization step S43 is a total treatment plan including one sub-portion (i.e., one BT sub-portion and one EBRT sub-portion) for each radiation set.

[0065] The BT sub-portion of the treatment plan is then delivered to the patient in step S45 and the actual delivered dose from that delivery is determined or estimated in step S46. Preferably, the situation after BT delivery is evaluated and used to improve the EBRT sub-portion of the treatment plan, as outlined in the following steps.

[0066] In step S47 , an updated image of the patient after BT treatment is obtained to account for geometric changes that occurred during BT treatment.

[0067] In a subsequent step S48, the EBRT sub-portion of the treatment is re-optimized taking into account the delivered dose from step S45 and the new images obtained in step S47. The optimization problem in this case comprises an objective function expressed as the following equation (2):

[0068]

[0069] in is the delivered dose determined in step S45.

[0070] The output from step S48 is a new optimized EBRT treatment plan S49, which is preferably delivered to the patient.

[0071] If targeted Figure 2 and Figure 3 As discussed, steps S46 through S49 may be performed without first performing the preceding steps, that is, EBRT planning may proceed even without an initial combined planning of the two for a previously delivered BT treatment.

[0072] It will also be possible to create a plan in which the BT and EBRT fractions are not given as two consecutive sub-fractions, but rather the BT fractions are distributed between the EBRT fractions. In this type of treatment, one or more sub-fractions of the treatment that have not yet been delivered can be re-optimized taking into account the delivered dose. Both the delivered and the non-delivered sub-fractions of the treatment will typically be a combination of BT and EBRT. The optimization in this case involves an objective function expressed as Equation (7):

[0073]

[0074] In all of the above approaches, simultaneous optimization should be performed carefully to ensure that the individual doses of each radiation set remain individually satisfactory. A possible adverse effect of co-optimization would be that the EBRT dose has cold spots in the target volume that are later filled by the BT dose. This can be mitigated by incorporating robustness into the model for, for example, uncertainties in instrument positioning and deformation effects of the instrument. Treatment-specific objective functions are also a possibility (similar to current beam set-specific objective functions).

[0075] As with any radiotherapy treatment plan, there will be sources of uncertainty, including patient placement, positioning of the BT device, and the estimated delivered dose. To compensate for this, robust planning can be used. In particular, deformation between images will result in an approximate cumulative dose, the quality of which depends on the accuracy of the deformable registration. To avoid over-optimization of a cumulative dose that differs from the dose that will actually be delivered, methods for robust planning based on the representation of uncertainty can be employed. Different degrees of refinement can be used, for example:

[0076] • Margin can be applied as an ITV on the predicted image, or as a smearing of the area to be treated only.

[0077] • Robust planning using scenarios generated as rigid shifts of the patient geometry can be applied independently to each of the radiation sets.

[0078] Robust plans using scenarios generated by multiple deformable registrations can be applied. This would involve utilizing multiple predictions in the case of delivering an EBRT fractional dose before a BT fractional dose. This would involve or utilize perturbations of the registration between acquired images in the case of delivering a BT fractional dose before an EBRT fractional dose. Deformations can also result from anatomical changes during EBRT, such as tumor shrinkage.

[0079] The method according to an embodiment of the present invention can also be combined with multi-criteria optimization. In this case, navigation can be performed in several aspects, with several trade-off objectives for either the total dose or the individual treatment doses.

[0080] Figure 5 is a schematic representation of a computer system that can perform the method of the present invention. The computer 31 comprises a processor 33 connected to a first data memory 34 and a second data memory 35, as well as a program memory 36. Preferably, there are also one or more user input devices 38, 39 in the form of a keyboard, mouse, joystick, voice recognition device, or any other available user input device. The user input device may also be arranged to receive data from an external storage unit.

[0081] The first data memory 34 contains the necessary data for carrying out the method, such as necessary images. The second data memory 35 stores data related to one or more current patients for whom a treatment plan is to be prepared. The program memory 36 stores a computer program arranged to cause the computer to execute, for example, a combination of Figure 2 、 Figure 3 and Figure 4 Any of the method steps discussed in .

[0082] As will be appreciated, data stores 34, 35 and program store 36 are schematically shown and discussed. There may be several data stores, each storing one or more different types of data, or one data store storing all data in a suitably structured manner, as well as a program store. As long as the components are able to communicate with each other, one or more of these components may be found in a cloud environment.

Claims

1. A computer-based method of optimizing a radiation therapy treatment plan for delivering radiation therapy to a patient with a second set of radiation, the method comprising the steps of: a) determining a dose previously delivered to the patient by means of a first set of radiation based on at least a first medical image of the patient, b) obtaining a first optimization problem comprising an optimization function designed to optimize a dose distribution resulting from the second set of radiations according to a dose criterion set for a total dose and a previously delivered dose, the total dose being the dose resulting from the second set of radiations and the previously delivered set of radiations, c) optimizing the treatment plan for delivering radiation therapy by the second radiation set by means of the optimization problem, wherein the optimization function is an objective function or a constraint and one of the first radiation set and the second radiation set is brachytherapy and the other is external beam radiation therapy, the optimization comprising deforming at least one of the dose distribution resulting from the second radiation set and previously delivered doses to a common geometry and accumulating them using a biological model, the optimization function being a set of a penalty for the accumulated dose and a penalty for radiation set specific doses from the second radiation set.

2. The computer-based method of claim 1 , wherein: The optimization function is designed to optimize the dose distribution resulting from the second set of radiation to match a difference between a desired total dose distribution and a previously delivered dose.

3. A computer-based method according to claim 1 or 2, wherein: The dose criteria are set based on at least an earlier medical image acquired prior to delivery of the first set of radiation.

4. The computer-based method according to claim 1 or 2, further comprising the following steps before step a): d) obtaining an initial optimization problem including an objective function designed to optimize the total dose distribution as a combination of a first dose distribution to be provided by the first set of radiation and a second dose distribution to be provided by the second set of radiation; e) Optimizing the treatment plan as a combination of the first radiation set and the second radiation set by means of the initial optimization problem.

5. The computer-based method of claim 1 or 2, wherein: The first optimization problem is obtained based on the second medical image.

6. The method according to claim 5, wherein: The second set of radiation is brachytherapy, and the second medical image includes a brachytherapy device applied to the patient.

7. The method according to claim 1 or 2, wherein: The at least one first medical image comprises at least one image of the patient acquired after delivery of the first set of radiation.

8. The method according to claim 1 or 2, wherein: The at least one first image includes at least one simulated image based on an estimate of the geometry of the patient after delivery of the first set of radiation. 9 . The method of claim 8 , the step of deforming one of the dose distributions involving deforming the dose of the second set of radiation to the geometry of the first medical image and integrating the dose using a biological model.

10. The method according to claim 1 or 2, wherein: Robust planning is used to account for uncertainties in brachytherapy delivery, the EBRT delivery, and / or the determined delivered dose.

11. A computer program product comprising computer readable code means arranged, when run in a computer, to cause the computer to perform the method according to any one of claims 1 to 10.

12. A computer system comprising a processor and at least one program memory, characterized in that: The program memory stores the computer program product according to claim 11.

Images