LU500593B1 - Automatic direct aperture optimization method and system - Google Patents
Automatic direct aperture optimization method and system Download PDFInfo
- Publication number
- LU500593B1 LU500593B1 LU500593A LU500593A LU500593B1 LU 500593 B1 LU500593 B1 LU 500593B1 LU 500593 A LU500593 A LU 500593A LU 500593 A LU500593 A LU 500593A LU 500593 B1 LU500593 B1 LU 500593B1
- Authority
- LU
- Luxembourg
- Prior art keywords
- dosage
- anfis
- weighting factor
- prescription
- target function
- Prior art date
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
- A61N5/00—Radiation therapy
- A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
- A61N5/103—Treatment planning systems
- A61N5/1036—Leaf sequencing algorithms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Pathology (AREA)
- Mathematical Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Automation & Control Theory (AREA)
- Fuzzy Systems (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Animal Behavior & Ethology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Radiation-Therapy Devices (AREA)
Abstract
The present disclosure relates to an automatic direct aperture optimization (DAO) method and system. The method includes: acquiring historical clinical data; respectively training a weighting factor adaptive neuro-fuzzy inference system (ANFIS) and a prescription dosage ANFIS according to the historical clinical data by using a matrix laboratory (MatLab) fuzzy inference toolkit in combination with a hybrid learning algorithm, to determine a trained weighting factor ANFIS and a trained prescription dosage ANFIS; and automatically and iteratively adjusting a weighting factor and a prescription dosage of a target function used by DAO, to determine a radiotherapy plan meeting clinical dosage requirements. The present disclosure shortens the optimization time, and reduces the workload of the physicist.
Description
BL-5290
AUTOMATIC DIRECT APERTURE OPTIMIZATION METHOD AND SYSTEM HUS00593
[01] The present disclosure relates to the field of direct aperture optimization (DAO), and in particular to an automatic DAO method and system.
[02] As the common treatment for cancers, radiotherapy includes the two step progress (TSP) and the direct aperture optimization (DAO). The DAO overcomes the inherent drawbacks of the
TSP, considers constraints on hardware of the multi-leaf collimator (MLC) during implementation, and generates an aperture that can be directly executed.
[03] Existing DAO is principally to optimize properties of the aperture, and there is a need for the physicist to repeatedly and manually adjust a prescription dosage parameter and a weighting factor parameter in the target function, which increases the optimization time of the radiotherapy scheme and the workload of the physicist.
[04] An objective of the present disclosure is to provide an automatic DAO method and system, to solve the problems of long optimization time of the radiotherapy scheme and a heavy workload of the physicist.
[05] To implement the above objective, the present disclosure provides the following solutions:
[06] The present disclosure provides an automatic DAO method and system. The method includes: reading a large amount of acquired historical clinical practice data to serve as training data, and training a weighting factor adaptive neuro-fuzzy inference system (ANFIS) and a prescription dosage ANFIS with a hybrid learning algorithm of a back propagation algorithm and a least squares method to determine parameters in a system; and automatically and iteratively adjusting, upon completion of training the weighting factor ANFIS and the prescription dosage ANFIS, a weighting factor parameter and a prescription dosage parameter in a target function for DAO, to obtain a radiotherapy plan meeting clinical dosage requirements. The present disclosure can effectively and automatically determine the prescription dosage parameter and the weighting factor parameter for the target function used in the DAO, thereby implementing the automatic DAO, shortening the optimization time, and reducing the workload of the physicist.
[07] To describe the technical solutions in the embodiments of the present disclosure or in the 1
BL-5290 prior art more clearly, the accompanying drawings required for the embodiments are briefly described below. Apparently, the accompanying drawings in the following description show merely some embodiments of the present disclosure, and a person of ordinary skill in the art may still derive other accompanying drawings from these accompanying drawings without creative efforts.
[08] FIG 1 is a flow chart of an automatic DAO method provided by the present disclosure.
[09] FIG 2 is a flow chart of an automatic DAO method according to Embodiment 1.
[10] FIG 3 is a schematic view of steps during training and implementation according to the present disclosure.
[11] The present disclosure is further described in detail below with reference to accompanying drawings and specific embodiments.
[12] As shown in FIG 1, an automatic DAO method includes the following steps:
[13] Step 101: Obtain historical clinical data.
[14] Step 102: Respectively train a weighting factor ANFIS and a prescription dosage ANFIS according to the historical clinical data by using a matrix laboratory (MatLab) fuzzy inference toolkit in combination with a hybrid learning algorithm, to determine a trained weighting factor
ANFIS and a trained prescription dosage ANFIS.
[15] Step 103: Automatically and iteratively adjust, according to the trained weighting factor
ANFIS and the trained prescription dosage ANFIS, a weighting factor and a prescription dosage of a target function used by DAO, to determine a radiotherapy plan meeting clinical dosage requirements.
[16] The target function is composed of sub-target functions through weighted summation. The number of subsystems and the number of sub-target functions of the target function are the same.
Each subsystem is provided with a weighting factor data block and a prescription dosage data block.
[17] Training the weighting factor ANFIS (referred to as the weighting factor system) includes the following steps:
[18] Step A: Extract a weighting factor data block of a corresponding sub-target function in the historical clinical practice data as training data, where the training data serves as an input variable and an output variable for training the weighting factor system of the sub-target function, the input variable is a deviation percentage between an actual dosage and the prescription dosage, and the output variable is an adjustment amount of the weighting factor.
[19] Step B: Input the training data to the weighting factor system to obtain an output result of the weighting factor system, and adjust parameters of the weighting factor system based on a 2
BL-5290 parameter learning rule of back propagation in combination with the output result, where the parameters include a premise parameter and a conclusion parameter, the parameter adjustment includes an adjustment on a center, a width, a slope and the like relevant to a membership function, the premise parameter is determined by a shape of a membership function of a fuzzifier, and the conclusion parameter is determined by a shape of a membership function of an anti-fuzzifier.
[20] Step C: Calculate a least mean square error (LMSE) of the weighting factor system, end the training on the weighting factor system of the sub-target function if the value reaches to a preset threshold, or otherwise, execute a next step.
[21] Step D: Read a set of new training data, and re-execute Steps B to C.
[22] The training process of the prescription dosage ANFIS is the same as that of the weighting factor system.
[23] Determining the radiotherapy plan specifically includes the following steps:
[24] Step A: Image a lesion area of a patient through a radiotherapy hardware device to obtain three-dimensional density information of the patient.
[25] Step B: Delineate an organ on an image of the lesion area through a physicist or automatic delineation software to obtain delineation information of each organ of the patient.
[26] Step C: Input treatment head information and target function information, and initialize a prescription dosage of an organ in the target function and a weighting factor of each sub-target function of the target function.
[27] Step D: Determine, according to the delineation information of each organ, an isocenter and a beam’s eye view (BEV), a beam to be calculated in each direction.
[28] Step E: Calculate a dosage distribution of each beam according to the three-dimensional density information and the treatment head information through a dosage calculation engine, thereby obtaining a dosage-deposition matrix in different BEVs.
[29] Step F: Determine an initial shape of an aperture from an outline of a target area in a BEV, calculate a weight value of an existing aperture according to the target function information and the dosage-deposition matrix through an aperture weight optimization algorithm, delete an aperture having a weight value of zero, fine-tune shapes of rest apertures, determine whether the number of present apertures exceeds a preset upper limit value, and if yes, no longer add a new aperture during iteration of the algorithm, only optimize the shape of the aperture, the shape optimization of the aperture and the weight optimization of the aperture being carried out alternately, and output, after optimizing the above solution for a period of time or a certain number of iteration times, present optimized aperture shape and aperture weight.
[30] Step G: Calculate a dosage distribution of each organ and a value of each sub-target function according to the aperture shape and the aperture weight, evaluate whether a present plan is 3
BL-5290 the radiotherapy plan meeting the clinical dosage requirements, and if yes, output the radiotherapy plan, or otherwise, execute Step H.
[31] Step H: Automatically adjust a weighting factor of each sub-target function according to the trained weighting factor system, continuously optimize the solution of Step G according to an adjusted weight of the sub-target function, the solution optimization and the adjustment of the weighting factor ANFIS being performed circularly, evaluate, after a certain number of iteration times, whether a present plan is the radiotherapy plan meeting the clinical dosage requirements, and if yes, output the radiotherapy plan, or otherwise, execute Step I.
[32] Step I: Automatically adjust a prescription dosage of each sub-target function according to the trained prescription dosage ANFIS, execute Step D to Step H based on an initial prescription dosage parameter and an initial weighting factor of each sub-target function of the target function, and end the ANFIS-based DAO until a plan meeting the clinical dosage requirements is obtained or the maximum number of iteration times for the optimization reaches.
[33] The target function information includes a type and used parameter information of each sub-target function. The algorithm used by the dosage calculation engine is a pencil beam algorithm, a monte carlo algorithm or a point accounting algorithm or the like. The optimization algorithm for optimizing the aperture shape and the aperture weight is a deterministic optimization algorithm, a stochastic optimization algorithm, and a hybrid algorithm of the deterministic and stochastic optimization algorithms. The evaluation of the plan includes comprehensive evaluation on the target dosage coverage and the organ-at-risk (OAR) sparing.
[34] Outputting the radiotherapy plan includes: Input information such as the optimized aperture shape and aperture weight, and a radiation source to a radiotherapy device, where a shape of an opening of an MLC is controlled by the aperture shape, and radiation exposure time of the opening of the MLC is controlled by the aperture weight corresponding to the aperture shape; and output a three-dimensional dosage distribution optimized result and a dosage-volume curve optimized result to a computer to generate a report.
[35] Based on the technical solutions of the present disclosure, the present disclosure is described with the specific embodiment.
[36] Embodiment 1
[37] As shown in FIG 2, the method in the embodiment includes the following steps:
[38] Step 201: Set a target function.
[39] The target function used in the radiotherapy includes a physical target function, a biological target function and a physical and biological hybrid target function. The physical target function includes a physical criterion sub-target function, the biological target function includes a biological criterion sub-target function, and the physical and biological hybrid target function includes the 4
BL-5290 physical criterion sub-target function and the biological criterion sub-target function. The physical criterion sub-target function includes a maximal dosage sub-target function, a minimal dosage sub-target function, an average dosage sub-target function, a minimal dosage-volume sub-target function and a maximal dosage-volume sub-target function; and the biological criterion sub-target function includes an equivalent uniform dosage (EUD) sub-target function, a normal tissue complication probability (NTCP) sub-target function and a tumor control probability (TCP) sub-target function. Setting the target function includes: Select a type of the function and select initial parameters in the target function, where the initial parameters include a prescription dosage and a weighting factor.
[40] Step 202: Perform DAO based on a simulated annealing algorithm.
[41] During DAO, the number of apertures required in each BEV is given, and the shapes and intensities of the apertures are optimized at the same time with the simulated annealing algorithm.
The initial shape of the aperture depends on an outline of a target area on the BEV map.
[42] Step 203: Evaluate a radiotherapy plan.
[43] The quality evaluation is made on the optimized radiotherapy plan in terms of a three-dimensional dosage distribution, a dosage-volume histogram (DVH) curve, a biological criterion, etc. Through the three-dimensional dosage distribution, the dosage coverage on the target area, and high-dosage irradiation on surrounding normal tissues and OARs are observed. Whether the present plan meets the conventional clinical radiotherapy dosage-volume evaluation standards is determined by viewing the DVH curve.
[44] In each embodiment of the present disclosure, the dosage coverage with 95% of prescription dosage and 107% of prescription dosage on the clinically concerned target areas, and the dosage-volume points on the clinically concerned OARs are evaluated to determine whether the present plan is the plan meeting the clinical dosage requirements.
[45] Step 204: Determine whether the present radiotherapy plan is practicable.
[46] According to the plan evaluation result obtained in Step 203, if the present plan meets all plan evaluation standards, i.e., the clinical dosage requirements, execute Step 205 to output the plan, or otherwise, execute Step 206 to perform an adjustment by the weighting factor ANFIS.
[47] Step 205: Output the radiotherapy plan.
[48] Shapes and intensities of all apertures in each BEV are output.
[49] Step 206: Perform an adjustment by the weighting factor ANFIS.
[50] The weighting factor ANFIS outputs a change Wichange of the weighting factor parameter for each sub-target function. If the present sub-target function has a weight value of w,,, and the modified sub-target function has a weight value of w _ , there will be a following relationship therebetween:
BL-5290
LU500593
Wow = Wo | 1+ ide > W change
[51] i=1
[52] where, n represents the number of sub-target functions, and the ; represents an jth sub-target function. The weighting factor parameters of all sub-target functions are modified similarly. After the weighting factor parameter is adjusted, re-execute Steps 202 to 205 once.
[53] Step 207: Evaluate the radiotherapy plan.
[54] Step 208: Determine whether the present radiotherapy plan meets the clinical dosage requirements.
[55] Step 208 and Step 204 are functionally identical. According to the plan evaluation result obtained in Step 207, if the present plan meets all plan evaluation standards, i.e., the clinical dosage requirements, execute Step 205 to output the plan, or otherwise, execute Step 209 to perform an adjustment by the prescription dosage ANFIS.
[56] Step 209: Perform an adjustment by the prescription dosage ANFIS.
[57] The prescription dosage ANFIS outputs a change Dichanze of the prescription dosage for each sub-target function. If the present sub-target function has a prescription dosage of D,,, and the modified sub-target function has a prescription dosage of D,_ , there will be a following relationship therebetween:
De, =D, +
DD ange
[58] =!
[59] where, n represents the number of sub-target functions, and the ; represents an jth sub-target function. The prescription dosage parameters of all sub-target functions are modified similarly. After the prescription dosage parameter in the target function is adjusted, re-execute Steps 202 to 208.
[60] The ANFIS-based automatic DAO method is performed iteratively until the plan meeting the clinical dosage requirements is generated, or the optimization is ended after the maximum number of iteration times, and the shapes and intensities of all apertures in each BEV are output.
[61] All contents in the above embodiment may be applied to the automatic DAO system disclosed in the present disclosure.
[62] As shown in FIG 3, the automatic DAO of the present disclosure is an iterative process.
During iteration, except that the shape and intensity of the aperture are automatically optimized, the prescription dosage parameter and the weighting factor parameter in the target function are 6
BL-5290 automatically determined for the DAO without the reliance on the manual adjustment of The >% physicist.
[63] By training the weighting factor ANFIS and the prescription dosage ANFIS according to the historical clinical data on parameter adjustment experience, and incorporating the parameter adjustment and inference experience of the physicist into the automatic DAO process, the present disclosure can effectively determine the prescription dosage parameter and the weighting factor parameter for the target function used in the DAO, and improves the efficiency of radiotherapy. 7
Claims (4)
1. An automatic direct aperture optimization (DAO) method, comprising: acquiring historical clinical data, the historical clinical data comprising a weighting factor data block and a prescription dosage data block; respectively training a weighting factor adaptive neuro-fuzzy inference system (ANFIS) and a prescription dosage ANFIS according to the historical clinical data by using a matrix laboratory (MatLab) fuzzy inference toolkit in combination with a hybrid learning algorithm, to determine a trained weighting factor ANFIS and a trained prescription dosage ANFIS, the hybrid learning algorithm comprising a back propagation algorithm and a least squares method; and automatically and iteratively adjusting, according to the trained weighting factor ANFIS and the trained prescription dosage ANFIS, a weighting factor and a prescription dosage of a target function used by DAO, to determine a radiotherapy plan meeting clinical dosage requirements.
2. The automatic DAO method according to claim 1, wherein the respectively training a weighting factor ANFIS and a prescription dosage ANFIS according to the historical clinical data by using a MatLab fuzzy inference toolkit in combination with a hybrid learning algorithm, to determine a trained weighting factor ANFIS and a trained prescription dosage ANFIS specifically comprises: enabling the weighting factor data block to comprise a deviation percentage between an actual dosage and the prescription dosage, as well as an adjustment amount of the weighting factor, determining an output result with the deviation percentage between the actual dosage and the prescription dosage as an input and the adjustment amount of the weighting factor as an output; adjusting parameters based on the back propagation algorithm in combination with the output result, to determine an adjusted weighting factor ANFIS, the parameters being parameters relevant to a membership function of a fuzzifier, and comprising a center, a width and a slope; calculating a minimum mean square error (MMSE) of the adjusted weighting factor ANFIS to determine the trained weighting factor ANFIS; enabling the prescription dosage data block to comprise the deviation percentage between the actual dosage and the prescription dosage, as well as an adjustment amount of the prescription dosage; determining an output result with the deviation percentage between the actual dosage and the prescription dosage as an input and the adjustment amount of the prescription dosage as an output; adjusting parameters based on the back propagation algorithm in combination with the output result, to determine an adjusted prescription dosage ANFIS; and calculating an MMSE of the adjusted prescription dosage ANFIS to determine the trained 8
BL-5290 prescription dosage ANFIS. 500595
3. The automatic DAO method according to claim 1, wherein the determining a radiotherapy plan meeting clinical dosage requirements specifically comprises: imaging a lesion area of a patient, to determine a lesion imaging area and acquire three-dimensional density information on a lesion of the patient; delineating an organ for the lesion imaging area to determine delineation information of each organ; acquiring treatment head information and the target function, and initializing the prescription dosage used in the target function and a a weighting factor of each sub-target function of the target function; determining, according to the delineation information, a beam to be calculated in each direction; calculating, according to the three-dimensional density information and the treatment head information through a dosage calculation engine, a dosage distribution of each beam to be calculated, thereby obtaining a dosage-deposition matrix in different beam's eye views (BEVs); determining an initial shape of an aperture from an outline of a target area in a BEV, and calculating a weight value of the aperture according to target function information and the dosage-deposition matrix through an aperture weight optimization algorithm; deleting an aperture having a weight value of zero, adjusting shapes of rest apertures, determining whether the number of present apertures exceeds a preset upper limit value, and if yes, no longer adding a new aperture during iteration, only optimizing the shape of the aperture, shape optimization of the aperture and the weight optimization of the aperture being carried out alternately, and outputting, after performing the optimization for set time or the set number of iteration times, present optimized aperture shape and aperture weight; calculating a dosage distribution of each organ and a value of each sub-target function according to the aperture shape and the aperture weight, evaluating whether a present plan is the radiotherapy plan meeting the clinical dosage requirements, and if yes, ending the optimization, and outputting the radiotherapy plan; and if no, automatically adjusting a weighting factor of each sub-target function according to the trained weighting factor ANFIS, and performing the optimization continuously according to an adjusted weight of the sub-target function, until the present plan is the radiotherapy plan meeting the clinical dosage requirements; and automatically adjusting, in case of still no radiotherapy plan meeting the clinical dosage requirements, a prescription dosage of each sub-target function according to the trained prescription dosage ANFIS, performing the optimization continuously based on an initialized prescription 9
BL-5290 dosage parameter and an initialized weighting factor, until a radiotherapy plan meeting the clinical > dosage requirements is obtained or the maximum number of iteration times for the optimization reaches.
4. An automatic direct aperture optimization (DAO) system, comprising: a data acquisition module, configured to acquire historical clinical data, the historical clinical data comprising a weighting factor data block and a prescription dosage data block; a training module, configured to respectively train a weighting factor adaptive neuro-fuzzy inference system (ANFIS) and a prescription dosage ANFIS according to the historical clinical data by using a matrix laboratory (MatLab) fuzzy inference toolkit in combination with a hybrid learning algorithm, to determine a trained weighting factor ANFIS and a trained prescription dosage ANFIS, the hybrid learning algorithm comprising a back propagation algorithm and a least squares method; and a radiotherapy plan determination module, configured to automatically and iteratively adjust, according to the trained weighting factor ANFIS and the trained prescription dosage ANFIS, a weighting factor and a prescription dosage of a target function used by DAO, to determine a radiotherapy plan meeting clinical dosage requirements.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| LU500593A LU500593B1 (en) | 2021-08-30 | 2021-08-30 | Automatic direct aperture optimization method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| LU500593A LU500593B1 (en) | 2021-08-30 | 2021-08-30 | Automatic direct aperture optimization method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| LU500593B1 true LU500593B1 (en) | 2023-02-28 |
Family
ID=85330674
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| LU500593A LU500593B1 (en) | 2021-08-30 | 2021-08-30 | Automatic direct aperture optimization method and system |
Country Status (1)
| Country | Link |
|---|---|
| LU (1) | LU500593B1 (en) |
-
2021
- 2021-08-30 LU LU500593A patent/LU500593B1/en active IP Right Grant
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110211664B (en) | System for automatically designing radiotherapy scheme based on machine learning | |
| US10850122B2 (en) | Optimization methods for radiation therapy planning | |
| US9555265B2 (en) | Intensity-modulated ion therapy | |
| CN110124214B (en) | Predictive dose distribution guided intensity modulated radiotherapy plan optimization systems, media and devices | |
| CN107441637A (en) | The intensity modulated radiation therapy Forecasting Methodology of 3-dimensional dose distribution and its application in the works | |
| CN114681813B (en) | Automatic radiation therapy planning system, automatic radiation therapy planning method, and storage medium | |
| US11116995B2 (en) | Radiation treatment planning based on dose rate | |
| CN110709134A (en) | Radiation therapy planning and dose aspects of treatment | |
| US11534625B2 (en) | Radiation treatment based on dose rate | |
| CN110327554B (en) | Intensity modulated radiotherapy plan optimization method based on predicted dose distribution guidance and application | |
| CN107823806B (en) | A kind of method and system for the direct Ziye optimization of intensity-modulated radiation therapy | |
| CN113557057A (en) | Graphical display of dose rate information for radiation treatment planning | |
| CN116209499A (en) | Dose and dose rate information correlation with volume for radiation treatment planning | |
| Zhang et al. | Plug pattern optimization for gamma knife radiosurgery treatment planning | |
| LU500593B1 (en) | Automatic direct aperture optimization method and system | |
| EP4252840A1 (en) | Monitor unit optimization constraint in radiation treatment planning | |
| Lu et al. | Reduced-order parameter optimization for simplifying prostate IMRT planning | |
| US20230191153A1 (en) | Assessing treatment parameters for radiation treatment planning | |
| CN113178242B (en) | Automatic plan optimization system based on coupled generation countermeasure network | |
| CN116341182A (en) | Proton intensity-modulated multi-objective optimization method and system based on cooperation of multi-objective evolution and traditional optimization method | |
| EP4106867B1 (en) | Generating and applying robust dose prediction models | |
| CN113870976A (en) | Dose pre-evaluation-based adaptive radiotherapy dose intensity modulation optimization calculation method | |
| CN110289075A (en) | A kind of direct Ziye optimization method and system based on fuzzy entropy | |
| Gutierrez et al. | A Reduced Variable Neighbourhood Search for the Beam Angle Optimisation Problem | |
| US11052265B2 (en) | Fluence map optimization for field-in-field radiation therapy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FG | Patent granted |
Effective date: 20230228 |