阅读说明：本技术 对放射疗法治疗进行计划的系统和方法 (The system and method that radiation therapy treatment is planned ) 是由 拉斯穆斯·博克兰茨 阿尔宾·弗雷德里克松 于 2017-01-27 设计创作，主要内容包括：系统(100)基于以下形式的输入对目标体积的放射疗法治疗进行计划：候选束的集合<Image he="70" wi="116" file="DDA0002131765080000011.GIF" imgContent="drawing" imgFormat="GIF" orientation="portrait" inline="no"></Image>其中每个束限定治疗束相对于目标体积的设置；用于放射疗法治疗的治疗计划(x),该治疗计划使用候选束的子集合<Image he="71" wi="127" file="DDA0002131765080000012.GIF" imgContent="drawing" imgFormat="GIF" orientation="portrait" inline="no"></Image>描述治疗计划(x)的质量的目标函数(F)；以及描述必须被满足的关于治疗计划(x)要求的可行域(X)。目标函数(F)和/或可行域(X)也反映将第一复杂度测量(τ(x))限制为小于或等于最大的第一复杂度(f)的第一复杂度准则(r(x)≤τ)。重复执行优化步骤；由此,在每次迭代中,通过根据目标函数(F)和可行域(X)对治疗计划(x)进行优化来计算更新的治疗计划(x’)。在此,如果满足终止准则,则基于更新的治疗计划(x’)计算选出束的集合(r’(x)≤τ’)。选出束的集合(r(x)≤τ)是候选束的集合的子集合。(The radiation therapy treatment of target volume is planned in the input of system (100) based on following form: the set of candidate beam Wherein each beam limits setting of the treatment beam relative to target volume；For the treatment plan (x) of radiation therapy treatment, which uses the subclass of candidate beam The objective function (F) of the quality for the treatment of plan (x) is described；And the feasible zone (X) required about treatment plan (x) that description must be satisfied.Objective function (F) and/or feasible zone (X), which also reflect, is limited to less than the first complexity measure (τ (x)) or is equal to the first complexity criterion (r (x)≤τ) of maximum first complexity (f).Repeat Optimization Steps；As a result, in each iteration, the treatment plan of update (x') is calculated by optimizing according to objective function (F) and feasible zone (X) to treatment plan (x).Here, the treatment plan (x') based on update calculates set (the r'(x)≤τ ' for selecting beam if meeting stop criterion).Select beam set (r (x)≤τ) be candidate beam set subclass.)

1. a kind of system (100) for being planned to radiation therapy treatment, in the radiation therapy treatment, target volume It is irradiated by treatment beam, the system (100) includes:

Processor (110), and

Memory (140), the memory (140) include the instruction (145) that can be executed by the processor (110), thus institute System (100) operation is stated into step coherent below executing:

[S₁₁] input set is obtained, the input set includes:

The set of candidate beamThe set of candidate's beamEach of candidate beam limit the treatment beam relative to The setting of the target volume；

For the treatment plan (x) of the radiation therapy treatment, the treatment plan (x) uses the subclass of the candidate beam

Objective function (F), the objective function (F) describe the quality of the treatment plan (x)；And

Feasible zone (X), the feasible zone (X) describe the requirement about the treatment plan (x) that must be satisfied,

At least one of the objective function (F) and the feasible zone (X) reflect the first complexity criterionIt is described First complexity criterionFirst complexity measure (τ (x)) is limited to less than to or is equal to maximum first complexityFirst complexity measure (τ (x)) is related to for the time of delivery of the treatment plan (x)；

[S₁₂] execute Optimization Steps, from there through according to the objective function (F) and the feasible zone (X) to the treatment plan (x) it optimizes and is calculated come the treatment plan (x') to update；And

[S₁₃] check whether and meet stop criterion；

And if it is satisfied, then [S₁₃₁] based on the treatment plan (x') of the update come to the set for selecting beamIt carries out It calculates, the set for selecting beamIt is the set of the candidate beamSubclass；

And if conditions are not met, then [S₁₃₂] treatment plan (x') of the update is set as the treatment plan (x)；To more The first new complexity criterionIt is calculated；By the first complexity criterion of the updateSetting For first complexity criterionAnd return to step [S₁₂]。

2. system according to claim 1 (100), wherein the system (100) further include:

At least one input interface (120), at least one described input interface (120) are configured to receive the objective function (F) With the restriction of the feasible zone (X), and

Output interface (130), the output interface (130) are configured to export through the determining set for selecting beam

3. according to claim 1 with system described in any one of 2 (100), wherein first complexity criterionIt indicates to the second complexity criterionIt is continuous and differentiable approach, second complexity criterionSecond complexity measure (t (x)) is limited in maximum second complexityInterior, second complexity is surveyed Measuring (t (x)) indicates the set of the candidate beam used in the treatment plan (x)In beam quantity, Yi Jineng The system (100) is set to be further operative to returning to step by the described instruction (145) that the processor (110) execute [S₁₂] before to the second complexity criterion of update Carry out calculating and the second complexity standard by the update ThenIt is set as second complexity criterion

4. system (100) according to any one of the preceding claims, wherein the input set further includes final Maximum second complexity (t_max), and the system can be made by the described instruction (145) that the processor (110) execute (100) the first complexity criterion of the further operating update in pairsIt is calculated, so that returning to step Suddenly [S₁₂] final iteration before, maximum first complexityEqual to final maximum first complexity (τ_max)。

5. system (100) according to claim 3, wherein the input set further includes final maximum second multiple Miscellaneous degree (t_max), and the system (100) further operating can be made by the described instruction (145) that the processor (110) execute Second complexity criterion of the pairs of updateIt is calculated, so that returning to step [S₁₂] it is final Before iteration, maximum second complexityEqual to the final maximum second complexity (t_max)。

6. system (100) according to claim 5, wherein the stop criterion is to be at least one of following item Basis:

[a] step [S₁₂] maximum number of iterations,

[b] described first complexity measure (τ (x)) is less than or equal to the final maximum first complexity (τ_max),

[c] described second complexity measure (t (x)) is less than or equal to the final maximum second complexity (t_max), and

[d] receives user and inputs (S), and user's input (S) is specified to be ceased and desisted order.

7. system (100) according to any one of the preceding claims, wherein can be executed by the processor (110) Described instruction (145) be further operative to the system (100) back to step [S₁₂] renewal process, institute are executed before Stating renewal process includes:

Set of the treatment plan (x') based on the update to the candidate beam of updateIt is calculated, the time of the update Select the set of beamIt is the set of the candidate beamSubclass；

By the set of the candidate beam of the updateIt is set as the set of the candidate beam

To the requirement for reflecting the feasible zone (X) and the only set of the candidate beamIn beam can be in the treatment The feasible zone (X') of the update of the requirement used in plan (x) is calculated；And

The feasible zone (X') of the update is set as the feasible zone (X).

8. system (100) according to claim 7, wherein the described instruction that can be executed by the processor (110) (145) it is further operative to the system (100):

After the renewal process and returning to step [S₁₂] before, treatment plan (x') based on the update and described Feasible zone (X) calculates the treatment plan (x ") further updated, and the feasible zone (X) includes the further update Treatment plan (x ").

9. the system (100) according to any one of claim 7 or 8, wherein can be executed by the processor (110) Described instruction (145) be further operative to the system (100) after the renewal process and back to step [S₁₂] before execute following steps in a step:

The treatment plan (x') of the update is set as the treatment plan (x)；Or

The treatment plan (x ") further updated is set as the treatment plan (x).

10. the system (100) according to any one of claims 7 to 9, wherein to the collection of the candidate beam of the update It closesIt is calculated, so that returning to step [S₁₂] meet at least one of following item before:

[i]: the first complexity measure (τ (x ")) of the treatment plan (x ") further updated is less than or equal to the maximum The first complexityAnd

[ii]: the second complexity measure (t (x ")) of the treatment plan (x ") further updated be less than or equal to it is described most The second big complexity

11. system (100) according to claim 8, wherein to the set of the candidate beam of the updateIt is counted It calculates, so that target function value (F (x ")) controlling as close possible to the update of the treatment plan (x ") further updated Treat the target function value (F (x')) of plan (x').

12. the system according to any one of claim 7 to 11 (100), wherein can be held by the processor (110) Capable described instruction (145) is further operative to the system (100):

Pass through the set from the candidate beamMiddle at least one beam of exclusion carrys out the set to the candidate beam of the update It is calculated, described at least one beam being excluded contributes the treatment plan (x') of the update minimum.

13. system (100) according to any one of the preceding claims, wherein the objective function (F) and it is described can At least one of row domain (X) is according to intended dose (D) and desired radiological dose (D_ref) deviation treated to describe the radiation The quality of the plan of method treatment, the intended dose (D) are based on the image data (D for limiting the target volume_img) calculate 's.

14. system (100) according to any one of the preceding claims, wherein obtained in the input set The treatment plan (x) indicates one of following item:

Use the set of the candidate beamIn the subclass of the beam irradiate the target volume, candidate's beam SetThe subclass in the treatment beam of each beam be consistent with the target volume and there is uniform fluence Distribution, the set for the candidate beamThe subclass in each beam the treatment beam the fluence distribution It is to be arranged by hierarchical sequence, so that the average intended dose (D) for the target volume is equal to defined target dose；

Irradiation is not conveyed；Or

According to the treatment plan being previously generated, the set of the candidate beam is usedIn the subclass of the beam convey photograph It penetrates.

15. system (100) according to any one of the preceding claims, wherein the set of candidate's beamIt is institute The discrete of the continuous collection of the possibility setting for the treatment of beam is stated, the continuous collection of the possible setting is at least one in following item Based on person:

Multiple machine capabilities for the treatment of machine；

Positioning of the patient relative to treatment machine；And

It is expected that the specific requirement of setting.

16. the method that a kind of pair of radiation therapy treatment is planned, in the radiation therapy treatment, target volume is by treatment beam Irradiation, the method includes following coherent steps:

[S₁₁] input set is obtained in processor (110), the input set includes:

The set of candidate beamThe set of candidate's beamIn each beam limit the treatment beam relative to described The setting of target volume；

For the treatment plan (x) of the radiation therapy treatment, the treatment plan (x) uses the subclass of the candidate beam

Objective function (F), the objective function (F) describe the quality of the treatment plan (x)；And

Feasible zone (X), the feasible zone (X) describe the requirement about the treatment plan (x) that must be satisfied,

At least one of the objective function (F) and the feasible zone (X) reflect the first complexity criterionIt is described First complexity criterionFirst complexity measure (τ (x)) is limited to less than or is equal to maximum first and is answered Miscellaneous degreeFirst complexity measure (τ (x)) is related to for the time of delivery of the treatment plan (x)；

[S₁₂] execute Optimization Steps, from there through according to the objective function (F) and the feasible zone (X) to the treatment plan (x) it optimizes and is calculated come the treatment plan (x') to update；

[S₁₃] check whether and meet stop criterion；

And if it is satisfied, then [S₁₃₁] based on the treatment plan (x') of the update come to the set for selecting beamIt carries out It calculates, the set for selecting beamIt is the set of the candidate beamSubclass；

And if conditions are not met, then [S₁₃₂] treatment plan (x') of the update is set as the treatment plan (x)；To more The first new complexity criterionIt is calculated；By the first complexity criterion of the updateSetting For first complexity criterionAnd return to step [S₁₂]。

17. according to the method for claim 16, further includes:

The restriction of the objective function (F) and the feasible zone (X) is received by least one input interface (120), and

It is exported by output interface (130) through the determining set for selecting beam

18. method described in any one of 6 or 17 according to claim 1, wherein first complexity criterion It indicates to the second complexity criterionIt is continuous and differentiable approach, second complexity criterionSecond complexity measure (t (x)) is limited in maximum second complexityInterior, second complexity is surveyed Measuring (t (x)) indicates the set of the candidate beam used in the treatment plan (x)In beam quantity, and institute State method further include:

Returning to step [S₁₂] before, to the second complexity criterion of updateIt is calculated, and

By the second complexity criterion of the updateIt is set as second complexity criterion

19. method described in any one of 6 to 18 according to claim 1, wherein input set further include it is final most The first big complexity (τ_max), and the method also includes:

To the first complexity criterion of the updateIt is calculated, so that returning to step [S₁₂] it is final Before iteration, maximum first complexityEqual to the final maximum first complexity (τ_max)。

20. according to the method for claim 18, wherein the input set further includes final maximum second complexity (t_max), and the method also includes:

To the second complexity criterion of the updateIt is calculated, so that returning to step [S₁₂] it is final Before iteration, maximum second complexityEqual to the final maximum second complexity (t_max)。

21. according to the method for claim 20, wherein the stop criterion is based at least one of following item :

[a] step [S₁₂] maximum number of iterations,

[b] described first complexity measure (τ (x)) is less than or equal to the final maximum first complexity (τ_max),

[c] described second complexity measure (t (x)) is less than or equal to the final maximum second complexity (t_max), and

[d] receives user and inputs (S), and user's input (S) is specified to be ceased and desisted order.

22. method described in any one of 6 to 21 according to claim 1, further includes: returning to step [S₁₂] before, it holds The renewal process that row includes the following steps:

Based on the treatment plan (χ ') of the update come the set of the candidate beam to updateIt is calculated, the time of the update Select the set of beamIt is the set of the candidate beamSubclass；

By the set of the candidate beam of the updateIt is set as the set of the candidate beam

To the requirement for reflecting the feasible zone (X) and the only set of candidate's beamIn beam can be counted in the treatment The feasible zone (X') for drawing the update of the requirement used in (x) is calculated；

The feasible zone (X') of the update is set as the feasible zone (X).

23. according to the method for claim 22, further includes:

After the renewal process and returning to step [S₁₂] before, treatment plan (x') based on the update and described Feasible zone (X) calculates the treatment plan (x ") further updated, and the feasible zone (X) includes the further update Treatment plan (x ").

24. the method according to any one of claim 22 and 23, further includes: after the renewal process and returning Return to step [S₁₂] before, execute a step in following steps:

The treatment plan (x') of the update is set as the treatment plan (x)；Or

The treatment plan (x ") further updated is set as the treatment plan (x).

25. the method according to any one of claim 22 to 24, comprising: to the set of the candidate beam of the updateIt is calculated, so that returning to step [S₁₂] meet at least one of following item before:

[i]: the first complexity measure (τ (x ")) of the treatment plan (x ") further updated is less than or equal to the maximum The first complexityAnd

[ii]: the second complexity measure (t (x ")) of the treatment plan (x ") further updated be less than or equal to it is described most The second big complexity

26. according to the method for claim 25, comprising:

To the set of the candidate beam of the updateIt is calculated, so that the mesh of the treatment plan (x ") further updated Target function value (F (x')) of the offer of tender numerical value (F (x ")) as close possible to the treatment plan (x') of the update.

27. the method according to any one of claim 21 to 26, further includes:

Pass through the set from the candidate beamMiddle at least one beam of exclusion carrys out the set to the candidate beam of the update It is calculated, described at least one beam being excluded contributes the treatment plan (x') of the update minimum.

28. method described in any one of 6 to 27 according to claim 1, wherein the objective function (F) and described feasible At least one of domain (X) is according to intended dose (D) and desired radiological dose (D_ref) deviation the radiotherapy described The quality of the plan for the treatment of, and the method also includes:

Based on the image data (D for limiting the target volume_img) calculate the intended dose (D).

29. method described in any one of 6 to 28 according to claim 1, wherein described in being obtained in input set Treatment plan (x) indicates one of following item:

Use the set of the candidate beamIn the subclass of the beam irradiate the target volume, candidate's beam SetThe subclass in the treatment beam of each beam be consistent with the target volume and there is uniform fluence Distribution, the set for the candidate beamThe subclass in each beam the treatment beam the fluence distribution It is to be arranged by hierarchical sequence, so that the average intended dose (D) for the target volume is equal to defined target dose；

Irradiation is not conveyed；Or

According to the treatment plan being previously generated, the set of the candidate beam is usedIn the subclass of the beam convey photograph It penetrates.

30. method described in any one of 6 to 29 according to claim 1, wherein the set of candidate's beamIt is described The discrete of the continuous collection of the possibility setting of beam is treated, the continuous collection of the possible setting is at least one of following item Based on:

Multiple machine capabilities for the treatment of machine；

Positioning of the patient relative to treatment machine；And

It is expected that the specific requirement of setting.

31. a kind of computer program (145) including following instructions: when executing the finger at least one processor (110) When enabling, at least one described processor (110) is made to execute method described in any one of 6 to 30 according to claim 1.

32. a kind of non-volatile data carriers (140), the non-volatile data carriers (140) include previous claim institute The computer program (145) stated.

Technical field

This patent disclosure relates generally to the plans of radiation therapy treatment.More particularly, the present invention relate to control radiotherapy Treat the system and corresponding method planned.The invention further relates to a kind of computer programs and a kind of readable Jie of processor Matter.

Background technique

In radiation therapy treatment, dosage is transported to the treatment volume of patient using one or more treatments with radiation beam In.Here, the orientation of corresponding treatment radiation beam is important parameter.The orientation with radiation beam is treated usually by the position of radioactive source It sets, the rotation of the direction of central beam axis and treatment head determines.

In CyberKnife^TMIn machine, it can be directly controlled and be taken by the linear accelerator being mounted in robot arms To parameter.Here, the orientation for belonging to single source position, which is referred to as, belongs to common node.

In the machinery with the so-called C-arm for carrying linear accelerator, wherein radioactive source is around patient with constant The rotation of source axial line distance, beam are orientated three groups of the rotation for describing instrument table, patient bed and treatment head by equal centers and respectively The angle (collimator angle) of member limits.For the machine of C-arm installation, coplanar treatment and general non-co-planar treatment are usual Difference be that in coplanar treatment, bed angle is constant for all beams.

The turntable similar with the linear accelerator of C-arm installation, which can be used, in treatment beam from Part Ther machine comes It is directed toward patient.

For being immovable conveying technology (such as in three-dimensional potential theory (3D- in light period interfascicular fasciculus orientation CRT), in electrostatic field intensity modulation radiotherapy (IMRT) and intensity modulated Part Ther (IMPT)), the set of beam orientation is direct Set corresponding to treatment beam.

The set of beam orientation can also be converted into the continuous path for electric arc therapy, and beam is irradiating in electric arc therapy Period is inswept patient, such as in volume modulation electric arc therapy (VMAT).

For example, by " shortest path " between identification beam orientation, making the mobile beam institute between the set of orientation In the sense that the time needed is minimum, the set that discrete beam is orientated is converted into continuous arc track.3D-CRT, I MRT and VMAT is the example of different Photon Thherapy technologies, they are exported using the linear accelerator of installation C-arm.IMPT can be from Fixed bunch conveying is conveyed by using turntable.

The general considerations for the subclass for selecting beam appropriate to be orientated from the biggish candidate collection that beam is orientated is not easy very much In parsing.

Now, there are many different methods to optimize beam orientation, such as so-called greedy method (greedy Since zero beam and then method), this method is iteratively added a beam at this moment.Here, must be in each iteration It executes repeatedly optimization --- to non-selected each candidate's mono- suboptimization of Shu Zhihang --- and because the quantity of candidate beam is logical Chang Feichang is big, so these calculating tend to be high disposal intensity.Under any circumstance, each iteration select beam all and be can be with Farthest improve the candidate beam of target function value.Therefore, greedy method cannot change places offer in short calculating time content Feasible treatment plan.A small amount of Optimized Iterative is used only by each subproblem, or is even used only from first time iteration Gradient information, it is possible to reduce calculate the time.However, this simplification is so that beam selection height is close, therefore quality is relatively low.

MIXED INTEGER programmed method indicates another alternative, wherein passing through the structure to all possible solution It is global optimum that change, which is enumerated beam orientations optimized problem analysis,.The calculating cost of MIXED INTEGER programmed method is very high, therefore It is infeasible for practical problem size.

Stochastic search methods indicate another example, wherein if Further aim functional value, exchange these Shu Bingbao at random Hold exchange.This method also tends to be to have processing intensity, or at least time-consuming.There are also meta-heuristic method, Such as simulated annealing and genetic algorithm, they improve current choosing using the combination of random and predefined search rule The set of beam out.Stochastic search methods and meta-heuristic method are also required to the calculating time grown very much, in order to search to most of Rope space is sampled.

One additional option is to use the local search approach for considering continuous beam angle space, and attempt by using by sensitivity Property analysis or the gradients of finite difference formulations improve target function value.However, local search approach is only useful to finely tuning, because Beam orientations optimized problem has many local minimums.

Finally, method of geometry analyzes the anatomical structure of patient there are also method of geometry, and calculate the quality point of candidate Shu Fangxiang Number is without executing any treatment plan optimization.Method of geometry can be very fast.But their accuracy is limited, because they Optimize independent of actual treatment plan.

In view of the above problems and shortcoming, therefore, universal clinical practice is manually selection beam orientation.Certainly, this It is far from ideal.

Summary of the invention

The purpose of the present invention is mitigating the above problem, and a kind for the treatment of plan solution is provided, which solves Scheme realizes user-friendly interface, and relative to need the process resource for the subclass for selecting beam to be orientated be it is effective, To be used in the candidate collection that is orientated from the bigger beam that potentially can be used.

According to an aspect of the present invention, the purpose by a kind of system for being planned to radiation therapy treatment come It realizes, target volume is irradiated by treatment beam in the radiation therapy treatment.The system includes processor and memory.Memory packet Containing the instruction that can be executed by the processor, thus the system operatio is to execute following coherent step.

Input set is obtained, the input set includes: the set of candidate beam, wherein in the set of candidate's beam Each beam defines setting of the treatment beam relative to target volume；For the treatment plan of radiation therapy treatment, the treatment plan Use the subclass of candidate beam；The objective function of the quality for the treatment of plan is described；And describe must be satisfied about controlling Treat the feasible zone of the requirement of plan；The objective function and/or feasible zone reflect the first complexity criterion, and described first is complicated First complexity measure is limited to less than or is equal to maximum first complexity by degree criterion.First complexity measure and treatment are counted The time of delivery drawn is related；

Optimization Steps are executed, treatment plan is optimized come to update from there through according to objective function and feasible zone Treatment plan calculated；And

It checks whether and meets stop criterion.If it is satisfied, then based on the treatment plan of the update come to the collection for selecting beam Conjunction is calculated, wherein the set for selecting beam is the subclass of the set of candidate beam.If conditions are not met, the treatment meter of the update Treatment plan will be set to by drawing；First complexity criterion of update is calculated；First complexity criterion of the update It is set to the first complexity criterion；And the program returns to the Optimization Steps for another an iteration.

The system is advantageous because in the slave treatment plan proposed the strategy of successive elimination beam produce it is a kind of for Become have the smaller calculating process for calculating intensity for each iteration.In addition, initial step has needed relatively low calculating Ability.Therefore, corporate plan processing efficient can be made in the time and in terms of calculating effect.

A kind of embodiment of this aspect according to the present invention, the system also include that at least one input interface and output connect Mouthful.At least one input interface is configured to for example receive objective function via manually user input, local datastore With the restriction of feasible zone, or the restriction from online resource reception objective function and feasible zone.Output interface is configured to export warp The determining set for selecting beam, for example, on display or storage device locally and/or remotely to present.These Interface is advantageous, because they provide the flexibility of height in terms of the destination of the area of origin of source data and result data.

Another embodiment in terms of this according to the present invention, the first complexity criterion indicate the company of the second complexity criterion It is continuous and differentiable approach.That is, the second complexity measure is limited in maximum second complexity by the second complexity criterion, and And second complexity measure instruction the set of candidate's beam used in treatment plan in beam quantity.However, such Anywhere the normally so-called step function of two complexity criterions, the step function are all non-differentiable.Therefore, it is mentioned The first complexity criterion (based on described continuous and differentiable approach) out essentially contributes to optimization and calculates.In addition, can be by The instruction that processor executes make system further operating with: it is quasi- to the second complexity of update back to before Optimization Steps It then carries out calculating and the second complexity criterion of the update is set as second complexity criterion.It means that similar In the first complexity criterion, second complexity criterion is also updated before optimization continues.

The another embodiment of this aspect according to the present invention, the input set are also multiple comprising final maximum first Miscellaneous degree.Can by processor execute instruction also make system operatio with return to Optimization Steps final iteration before to described The first complexity criterion updated is calculated.Maximum first complexity is equal to final maximum first complexity.Therefore, The complexity of obtained treatment plan can be known in advance.

The another embodiment of this aspect according to the present invention is similar to above, and input set is comprising finally most The second big complexity, and the system further operating can be made to answer the second of the update by the instruction that processor executes Miscellaneous degree criterion is calculated, so that maximum second complexity is equal to most before the final iteration back to Optimization Steps Whole maximum second complexity.

The other embodiment of this aspect according to the present invention, stop criterion are based on one or more of following: [a] is excellent Change the maximum number of iterations of step, [b] first complexity measure is less than or equal to final maximum first complexity, [c] the Two complexity measures are less than or equal to final maximum second complexity, and [d] receives user's input, and the user is defeated Enter to specify and cease and desist order.Therefore, the flexibility of height is obtained in terms of the condition for terminating optimization process.

Another embodiment of this aspect according to the present invention can be such that system is also returning to by the instruction that processor executes It operates before Optimization Steps to execute the renewal process comprised the steps of.Based on the treatment plan of the update come to update The set of candidate beam is calculated, wherein the set of the candidate beam of the update is the subclass of the set of candidate beam；It will be described The set of the candidate beam of update is set as the set of candidate beam；To reflection feasible zone requirement and only candidate beam set in Beam can be in the feasible zone of the update of requirement used in treatment plan；And the feasible zone of the update is set as feasible zone It is calculated.This means that treatment plan is only based on the contributive beam of irradiation actually to treatment volume.

The Additional embodiments of this aspect according to the present invention, can be by the instruction that processor executes so that system is further grasped Make with after above-mentioned renewal process and back to before Optimization Steps, treatment plan and feasible zone based on the update The treatment plan further updated is calculated.Here, feasible zone includes the treatment plan further updated.Therefore, Optimization can be continued based on the treatment plan further updated.

Another embodiment of this aspect according to the present invention, can be by the instruction that processor executes so that system is further grasped Make so that after the renewal process and before returning to Optimization Steps, the treatment plan of the update is set as by (a) The treatment plan, or (b) set the treatment plan further updated according to the version for being currently available that treatment plan For the treatment plan.

The another embodiment of this aspect according to the present invention, before returning to the Optimization Steps, to the update The set of candidate beam calculated so that: the first complexity measure of the treatment plan further updated is less than or waits It is less than or waits in the second complexity measure of maximum first complexity and/or the treatment plan further updated In maximum second complexity.This makes the complexity in entire optimization process keep moderate.

The additional embodiment of this aspect according to the present invention calculates the set of the candidate beam of the update, makes The target function value of the treatment plan further updated must be used for as close possible to the treatment plan for the update Target function value.Alternatively, by excluding at least one beam from the set of candidate beam come the candidate beam to the update Set is calculated.The one or more beams being excluded contribute the treatment plan of the update minimum.In both of these case Under, fast convergence is expected.

Another embodiment of this aspect according to the present invention, objective function and/or feasible zone are according to intended dose and phase The deviation of the radiological dose of prestige describes the quality of the plan for the radiation therapy treatment.Here, the intended dose is Based on what is for example calculated by system by the received image data for limiting target volume of input interface.Therefore, conventional CT Scanning or MRI data can be used for treatment plan.

The other embodiments of this respect according to the present invention, the initial treatment plan obtained in the input set indicate One of three kinds of alternative solutions below.

According to the first alternative solution, treatment plan is related to irradiating target using the subclass of the beam in the set of candidate beam Volume.The treatment beam of each beam in the subclass of the set of candidate's beam is consistent with target volume and has uniform fluence point Cloth.In addition, the fluence distribution of the treatment beam of each beam in the subclass of the set for candidate beam is still set by hierarchical sequence It sets, so that the average intended dose for target volume is equal to defined target dose.

According to the second alternative solution, treatment plan is related to not conveying irradiation completely；And according to third alternative solution, treatment Plan is related to conveying irradiation using the subclass of the beam in the set of candidate beam according to the treatment plan being previously generated.

The another embodiment of this aspect according to the present invention, the set of candidate's beam are to treat the possible setting of beam Discrete (discrete results) of continuous collection.In turn, the continuous collection of the possible setting is based on: multiple mechanical energy for the treatment of machine The specific requirement of power, patient relative to positioning, and/or the expectation setting for the treatment of machine.Therefore, can by therapy to be used and The suitable parameter of machine is weighted in the optimization process.

According to another aspect of the present invention, the purpose by a kind of method for being planned to radiation therapy treatment come It realizes, target volume is irradiated by treatment beam in the radiation therapy treatment.This method includes following coherent step.

Input set is obtained in the processor, and wherein the input set includes: the set of candidate beam, wherein each beam limits Determine to treat setting of the beam relative to target volume；The treatment plan of radiation therapy treatment, the treatment plan use the candidate The subclass of beam；The objective function of the quality for the treatment of plan is described；And describe to be satisfied wants about treatment plan The feasible zone asked.Objective function and/or feasible zone reflect the first complexity criterion, and first complexity criterion is by the first complexity Measurement is limited to less than or is equal to maximum first complexity.In turn, the first complexity measure and the conveying for being used for treatment plan Time correlation.In other words, the first relatively low complexity measure corresponds to counts with the associated treatment of relatively short time of delivery It draws, and vice versa；

Optimization Steps are executed, treatment plan is optimized come to update from there through according to objective function and feasible zone Treatment plan calculated；And

It checks whether and meets stop criterion.If it is satisfied, then based on the treatment plan of update come to select the set of beam into Row calculates, wherein the set for selecting beam is the subclass of the set of candidate beam.Otherwise (that is, if stop criterion does not meet), institute The treatment plan for stating update will be set to the treatment plan；First complexity criterion of update is calculated；It is described more The first new complexity criterion is set to first complexity criterion；And the process returns to the above Optimization Steps.

The system proposed according to reference it is described above, the advantages of this method and its be preferably implemented and trans- become bright It is aobvious.

According to another aspect of the invention, which passes through the meter that can be loaded into the memory of at least one processor Calculation machine program realizes, and including for running described program on at least one processor when executes the soft of the above method Part.

According to another aspect of the present invention, which is realized by the processor readable medium having program recorded thereon thereon, Wherein described program controls at least one processor to be loaded into Shi Zhihang at least one described processor in described program The above method.

Compared with known solution, of the invention one general advantage was considered at least some stage of algorithm All beams orientations that may be available in machine.The present invention, which also can be controlled easily, calculates the time.

Other advantage, beneficial aspects and application of the invention will become bright according to the following description and the appended claims It is aobvious.

Detailed description of the invention

The present invention, these preferred embodiments will be explained by means of preferred embodiment and in greater detail with reference to the attached drawings now It is disclosed as example.

Fig. 1 shows a kind of overview of the treatment planning systems of embodiment according to the present invention；

Fig. 2 a-b has been illustrated expression according to embodiment of the present invention and approaches the second complexity measure of expression Step function continuous and differentiable function the first complexity measure；And

Fig. 3 shows universal method according to the present invention by means of process.

Specific embodiment

It is an object of the present invention to from potentially useful to provide the candidate that the biggish beam of regulation therapy for the treatment of volume is orientated Set in, select the subclass of suitable beam orientation come using.

In order to achieve this goal, it is necessary to meet the standard requirements for the treatment of plan optimization.Therefore, selection description is treated The set of parameter is as optimized variable.Further, it has been determined that the feasible set of these variables, and formulated Quantified therapy meter Draw the objective function of quality.The example of variable for photon technology (such as 3D-CRT, static field IMRT and VMAT) is every picture Energy fluence (surface-element in beam plane), multi-diaphragm collimator leaf position, dosage rate and the time of delivery of every part of element, Or Shu Quanchong.The canonical variable of IMPT is every spot scan time.Feasible variables collection by reflection output method physical limit Constraint limited, and can also include to intended dose distribution constraint.Objective function for treatment plan would generally Punishment is made to the deviation between intended dose and required dosage.However, objective function is it is also contemplated that the treatment plan Other aspects, the geometry complexity of such as optimized collimator shape.Treatment plan person would generally select about following: make For the dosage and variable of a part of objective function which has require (requirement minimized the error), and as formulation feasible zone Constraint have which require (requirement that must be fully met).However, variable, the definite selection of feasible zone and objective function and root It is unrelated according to parsing scheme of the invention.

The finite aggregate of candidate beam orientation must also be limited.Such set can be given birth in a manner of full automation At, or created based on the input for the treatment of plan person.It can be used for limiting the possibility parameter of the finite aggregate of candidate beam orientation It is: whether should allows non-coplanar beam orientation, whether should allow variable angle of collimation and possible, bed and collimation The discrete levels of the set at angle.Certainly, it should the angle that can cause to collide between patient and treatment head is removed from candidate set The combination of degree.It is also such for the angle in the set for forbidding angle distributed by treatment plan person due to other reasons.

Fig. 1 shows the system for being planned to radiation therapy treatment according to embodiment of the present invention 100 overview, in the radiation therapy treatment, target volume is treated beam irradiation.System 100 includes processor 110 and deposits Reservoir 140.Memory 140 again comprising can by the instruction 145 (i.e. software) that processor 110 executes, thus the operation of system 100 to Coherent step [S is executed in such a way that possibility is duplicate₁₁]、[S₁₂] and [S₁₃].However, in description step [S₁₁]、[S₁₂] and [S₁₃] before, we are by the basic principle describing Shu Youhua in a manner of mathematics to explain proposed program behind.

IfIt isThe quantity of middle element.

In addition, settingThe vector of optimized variable, the vector by define withIn it is given Beam is orientated the subvector x of the relevant optimized variable of b_bComposition.

X is the set (i.e. above-mentioned feasible zone) of feasible variable vector.

F is using x as the objective function measured to plan quality, wherein lower value is preferred.

It is measurement x_bTo the component y of the contribution of wholistic therapy_bFunction, which is defined as So that: if x_bWith the corresponding value for the treatment of is not used in b, then y_b(x_b)=0, otherwise y_b(x_b)>0。

T is by y (x) as the function measured to plan complexity, and wherein t is defined such that it depends on y (x) Radix (i.e. the quantity of the nonzero element of the vector), the function representation is at card (y (x)).In the following, we will explain in detail respectively Itd is proposed the first complexity measure τ (x) and the second complexity measure t (x) are stated, to quantify the complexity of the plan.

The example y of possible measurement_bBe the integral dose (integral absorbed dose) of target volume is transported to from beam b, the beam of beam b is opened Open time or vector x_bA certain norm.The radix of radix or y that function t may, for example, be y adds a certain norm of y.

First example corresponds to the case where multiple static beams should be controlled, and the second example is total corresponding to that should control The case where treatment time, wherein radix item indicates the Time constant cost of additional static beam, and norm item is indicated in this way Beam the beam opening time.The possibility of t for electric arc therapy limits the subclass for being connection for the beam orientation for the treatment ofMost short feasible path length (or a certain of this path length approaches value).For CyberKnife^TMThe restriction of the t for the treatment of is such vector: wherein radix of first component to the quantity of the node for treatment It measures, and second component measures the radix of the quantity of the beam orientation for treatment.

Beam orientations optimized can be described as following problems now:

Wherein t is limited in a certain maximum acceptable complexity t_maxIn the range of, or

Wherein G is the penalty function of t (x),

Such as positive weights λ, G (t (x))=λ (max { t (x)-t_max, 0 }.

(P1) or the solution x* of (P2) limits the set that selected beam is orientatedWherein,

However, optimization problem (P1) and (P2) are very difficult to parse, because the radix of y must be controlled during optimization. Know that this general convex radix problem refers to the polynomial time problem (NP-hard) of uncertainty.

Therefore, according to the present invention, replace finding problem (P1) or (P2) within the actually calculating time using heuristic function Approximating Solutions.Heuristic function passes through continuous and differentiable function τ_εT is approached, is thus more suitable for optimizing.Here, subscript ε is indicated Control the parameter of approximation accuracy.

Heuristic function is also by beam from the set of candidate beamIn in turn move on to the subclass for selecting beamMore accurately It says, the set of candidate beamPassing through willIt is defined asIn select the subclass of beam to update, then with the subclass for selecting beamTo replace the set of previous candidate beam

Therefore removed beam orientation is those beam orientations for being confined to make treatment zero contribution, or is at least made The beam of insufficient contribution is orientated.

In short, this following form for providing (P1) approaches corresponding:

Wherein,It is the complexity limit,

And provide the following form of (P2) approaches corresponding:

By utilize the fact that the Continuous Approximation value of the radix of n member vector z can be obtained: by utilizing card (z) can To be expressed as the sum of step function s, according to the following formula:

And step function s can be approached by continuous function.Herein, it should be noted that when card () is applied to y,

Fig. 2 a and Fig. 2 b show the first examples for the card (y) in positive domain being smooth Continuous Approximation and second to show Example.

More precisely, Fig. 2 a indicates the sum of logical function according to the following formula:

And

Fig. 2 b indicates the ε member norm of y according to the following formula:

When ε is close to zero, the two approach value close to card (y), as shown in Figure 2 a and 2 b, respectively indicate (A1) and (A2) the inside summand in.

In addition, the optimization problem in (A1) and (A2) is continuous optimization problems, many general solvers are asked suitable for these Topic.

As will be described below, the example for the stop criterion of optimization is: current iteration x is in some sense Convergence, or have occurred that some form of time-out, for example, having reached maximum number of iterations.Initial point for optimization can To be preselected initial guess (for example, uniformly fluence is distributed scale, so that target dose averagely gets off equal to defined agent Amount).

In the above content, following procedure is applied according to the present invention.Specifically, the instruction 145 in memory 140 can be by It manages device 110 to execute, so that the operation of system 100 is to execute coherent step [S₁₁]、[S₁₂] and [S₁₃]。

In step [S₁₁] in, input set is obtained from the memory in system 100 or via one or more interfaces 120.

The input set includes the set of such candidate beamThe wherein set of candidate's beamIn each beam limit Setting of the treatment beam relative to target volume.The input set is also comprising the subclass using candidate beamRadiotherapy The treatment plan x for the treatment of.In addition, the objective function F of quality of the input set comprising description treatment plan x and description must The feasible zone X for the requirement about treatment plan x that must be satisfied.

The treatment plan x obtained in the input set can indicate the set using candidate beamIn beam subclass To irradiate target volume.The set of candidate's beamSubclass in the treatment beam of each beam be consistent and have with target volume Uniform fluence distribution.For the set of candidate beamSubclass in each beam treatment beam fluence distribution or by point Grade series setting, so that the average intended dose D of target volume is equal to defined target dose.

Alternatively, the treatment plan x obtained in input set can indicate not conveying irradiation；Or according to first previous existence At treatment plan and use the set of candidate beamIn beam subclass come convey irradiation.

Objective function F and/or feasible zone X reflects the first complexity criterionFirst complexity criterion is by first Complexity measure τ (x) is limited to less than or is equal to maximum first complexityFirst complexity measure τ (x) and treatment plan x Time of delivery it is related.This means that the first relatively low complexity measure is corresponding to associated with relatively short time of delivery Treatment plan, and vice versa.

Objective function F and/or feasible zone X is preferably also according to intended dose D and desired radiological dose D_refDeviation and Describe the quality of the plan for radiation therapy treatment.In turn, based on the image data D for limiting target volume_imgCome to plan Dosage D is calculated.Image data D_imgIt can for example indicate CT scan data or MRI data.

In order to reduce calculation amount, the set of candidate beamPreferably treat the discrete of the continuous collection of the possibility setting of beam. In turn, the continuous collection that may be arranged can be based on: multiple machine capabilities for the treatment of machine, patient determine relative to treatment machine The specific requirement of position, and/or expectation setting.

In step [S₁₁] after the step of [S₁₂] in, Optimization Steps are executed, thus processor 110 can be according to target letter Number F and feasible zone X optimizes treatment plan x to calculate the treatment plan x' of update.Hereafter step [S is executed₁₃]。

In step [S₁₃] in, it checks whether and meets stop criterion；And if it is satisfied, then in step [S₁₃₁] in selecting The set of beamIt is calculated.Treatment plan x' based on update is to the set for selecting beamIt is calculated.Select the set of beamIt is the set of candidate beamSubclass.That is, selecting the set of beamInclude the set than candidate beamLess beam or extremely Mostly with the set of candidate beamBeam as many.

If in step [S₁₃] in discovery be unsatisfactory for stop criterion, then then carry out step [S₁₃₂], in step [S₁₃₂] in, The treatment plan x' of update is set to treatment plan x.In step [S₁₃₂] in, the first complexity criterion of update The first complexity criterion for being calculated, and being updatedIt is set to the first complexity criterionHereafter, program Back to step [S₁₂] further to be optimized.

Preferably, system 100 further includes at least one input interface 120 and output interface 130.At least one described input Interface 120 is configured to receive the restriction of objective function F and feasible zone X, such as from local datastore 153 or from online resource 152 receive.Certainly, objective function F can also pass through keyboard 151, touch screen or similar input link with the restriction of feasible zone X It inputs and inputs via manual user.In addition, at least one described input interface 120 can be for example from online resource 152 or sheet Ground data storage 153 is for example via keyboard 151 or online resource 152 and/or the image data D of restriction target volume_imgTo receive To desired radiological dose D_refDescription.Output interface 130 is configured to export through the determining set for selecting beamSuch as with To present on display 161, to be stored in local datastore 163 and/or remote data resources 162.

As described above, the first complexity criterionPreferably indicate the second complexity criterionIt is continuous and It is differentiable to approach.That is, it is this continuous and it is differentiable approach simplify with substantially with basic NP-hard problem carried out Corresponding calculating is related to optimize relevant calculating.Similar to the first complexity criterionSecond complexity criterion Second complexity measure t (x) is limited in maximum second complexityIt is interior.Here, the second complexity measure t (x) instruction is being controlled Treat the set of candidate's beam used in plan xIn beam quantity.

In order to gradually improve treatment plan, the instruction 145 that can be executed by processor 110 makes system 100 also return to step Suddenly [S₁₂] before can be used to the second complexity criterion to updateIt is calculated, and complicated by the second of update Spend criterionIt is set as the second complexity criterion

In order to ensure the reasonable convergence of optimization, the input set is preferably also comprising final maximum first complexity τ_max, and the instruction 145 that can be executed by processor 110 makes 100 further operating of system with the first complexity to the update CriterionIt is calculated, so that returning to step [S₁₂] final iteration before, maximum first complexityDeng In final maximum first complexity τ_max。

Alternatively, the input set includes final maximum first complexity τ_max, however can be by processor 110 The instruction 145 of execution makes the operation of system 100 for transferring the second complexity criterion to the updateIt is calculated, So that returning to step [S₁₂] final iteration before, maximum second complexityIt is multiple equal to final maximum second Miscellaneous degree t_max。

A kind of embodiment according to the present invention, stop criterion are based on: step [S₁₂] maximum number of iterations, first complicated Degree measurement τ (x) is less than or equal to final maximum first complexity τ_max, the second complexity measure t (x) be less than or equal to most Whole maximum second complexity t_max, and/or the specified user to cease and desist order inputs S and is received.Therefore, optimization process can It, can also be in any point-in-time knot selected by operator (such as treatment plan person) to terminate when meeting scheduled criterion Beam.

Preferably, the instruction 145 that can be executed by processor 110 makes system 100 also return to step [S₁₂] operate before To execute renewal process.The renewal process further relates to following steps.

It is calculated according to set of the updated treatment plan x' to the candidate beam of update.Here, the time of the update Select the set of beamIt is the set of candidate beamSubclass.

Then, the set of the candidate beam of the updateIt is set to the set of candidate beamAnd to the feasible zone of update X' is calculated, and the feasible zone X' of the update reflects the requirement of feasible zone X and the set of candidate beamIn beam only can be with It is required used in treatment plan x.Hereafter, the feasible zone X' of update is set to feasible zone X, and program can be recycled back to To step [S₁₂] further to be optimized.

However, after above-mentioned renewal process and returning to step [S₁₂] before, processor 110 also preferably calculates The treatment plan x " for the treatment of plan x' and feasible zone X based on update further updated.Here, feasible zone X include it is described into The treatment plan x " that one step updates.

Furthermore it is preferred that calculating the set of the candidate beam of the updateSo that the treatment plan for further updating Target function value F (x') of the target function value F (x ") of x " as close possible to the treatment plan x' for the update.That is, this Accelerate the convergence of optimization.

In addition to this, it is advantageous to if calculating the set of the candidate beam of the updateIt is related to the set from candidate beam Middle at least one beam of exclusion, and at least one beam therefore excluded is that minimum contribution is made to the treatment plan x' of update At least one beam.

In addition, after above-mentioned renewal process and returning to step [S₁₂] before, processor 110 be preferably carried out with Lower step: the step for the treatment of plan x' of update is set as treatment plan x；Or by the treatment plan further updated X " is set as the step for the treatment of plan x.Therefore, the quality of the treatment plan generated by the final iteration of Optimization Steps improves.

A kind of embodiment according to the present invention calculates the set of the candidate beam of the updateSo that returning to step Suddenly [S₁₂] meet one or two of the following conditions condition before.

The first complexity measure τ (x ") of [i] described treatment plan x " further updated is less than or waits

In maximum first complexityAnd

The second complexity measure t (x ") of [ii] described treatment plan x " further updated be less than or

Equal to maximum second complexity

In order to summarize, and with reference to the flow chart in Fig. 3, we will now be described according to the present invention for plan radiation Therapy treats conventional method, and in the method, target volume is treated beam irradiation.

In first step S₁₁In, input set is obtained in processor 110.The input set includes such candidate beam SetWherein each beam defines setting of the treatment beam relative to target volume.The input set for radiating also comprising treating The treatment plan x of method treatment, treatment plan x use the subclass of candidate beamIn addition, the input set includes description treatment Plan the objective function F of the quality of x, and describes the feasible zone X for the requirement about treatment plan x that must be satisfied.Target Function F and/or feasible zone X reflects the first complexity criterionFirst complexity criterion is by the first complexity measure τ (x) it is limited to less than or is equal to maximum first complexityThe time of delivery of first complexity measure τ (x) and treatment plan x It is related.

Then, in step S₁₂In, execute Optimization Steps, from there through according to objective function F and feasible zone X to treatment plan X optimizes to calculate the treatment plan x' of update.

Hereafter, step S₁₃It checks whether and meets stop criterion；If it is satisfied, then then carrying out step S₁₃₁.Otherwise, program Proceed to step S₁₃₂, in step S₁₃₂In, the treatment plan x' of the update is set to treatment plan x.In step S₁₃₁In, The first complexity criterion updatedAlso it is calculated, and the first complexity criterion of the updateIt is set It is set to first complexity criterionThen, program is circulated back to step S₁₂Further to be optimized.

In step S₁₃₁In, based on the treatment plan x' of the update to the set for selecting beamIt is calculated.Select beam SetIt is the set of candidate beamSubclass.Then, EP (end of program), and radiotherapy can be presented to treatment plan person Treatment, which is based on the set for selecting beamWith the treatment plan x' of the update.

It can be controlled by means of processor by programming with reference to Fig. 3 all processing steps described and these steps Any subsequence.In addition, although including processor and at least one above with reference to the embodiments of the present invention that attached drawing describes The process executed in a processor, but therefore the present invention also extends to computer program, especially on carrier or in the carrier The computer program for being suitable for putting the invention into practice.The program can be source code, object code, code intermediate source, Any other form of such as object code of partial compilation form or the implementation suitable for process according to the present invention Object code.The program can be a part of operating system, be also possible to individual application program.Carrier can be can Carry any entity or device of the program.For example, carrier may include storage medium, such as flash memory, ROM (read-only storage Device), such as (erasable programmable is read-only to be deposited by DVD (digital video/generic diskette), CD (compact disk) or semiconductor ROM, EPROM Reservoir), EEPROM (electrically erasable programmable read-only memory) or magnetic recording media, such as floppy disk or hard disk.In addition, carrier Can be can transport vehicle, the electric signal that can be such as transmitted by cable or optical cable or by radio or by other means Or optical signal.When the signal that described program tool chemical conversion can directly be transmitted by cable or other devices or equipment, the load Body can be made of this cable or device or equipment.Alternatively, carrier can be the integrated circuit for being embedded with described program, The integrated circuit is adapted for carrying out or for executing correlated process.

When used in this manual, term "comprises/comprising" is for specifying the feature, integer, step or component Presence.However, the term does not exclude the presence of one or more supplementary features, integer, step or component or their group, or Increase one or more supplementary features, integer, step or component or their group.

The invention is not restricted to the embodiment, but can freely change within the scope of the claims.