CN115620870A - Automatic planning method and system based on dose prediction and parameter optimization - Google Patents

Automatic planning method and system based on dose prediction and parameter optimization Download PDF

Info

Publication number
CN115620870A
CN115620870A CN202211144237.8A CN202211144237A CN115620870A CN 115620870 A CN115620870 A CN 115620870A CN 202211144237 A CN202211144237 A CN 202211144237A CN 115620870 A CN115620870 A CN 115620870A
Authority
CN
China
Prior art keywords
particle
dose
result
objective function
risk
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211144237.8A
Other languages
Chinese (zh)
Inventor
林谦
李君利
邱睿
周京京
武祯
张辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to CN202211144237.8A priority Critical patent/CN115620870A/en
Publication of CN115620870A publication Critical patent/CN115620870A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/40—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
    • 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
    • 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
    • A61N5/1031—Treatment planning systems using a specific method of dose optimization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20092—Interactive image processing based on input by user
    • G06T2207/20104—Interactive definition of region of interest [ROI]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/03—Recognition of patterns in medical or anatomical images

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Public Health (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Radiology & Medical Imaging (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Epidemiology (AREA)
  • Pathology (AREA)
  • Primary Health Care (AREA)
  • Veterinary Medicine (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Quality & Reliability (AREA)
  • Genetics & Genomics (AREA)
  • Databases & Information Systems (AREA)
  • Multimedia (AREA)
  • Urology & Nephrology (AREA)
  • Physiology (AREA)
  • Surgery (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

The invention discloses an automatic planning method and system based on dose prediction and parameter optimization, wherein the method comprises the following steps: performing planned CT scanning on a patient to obtain a CT image, and obtaining organ delineation data according to the CT image; inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction; determining an objective function of an irradiation field distribution model based on organ delineation data and the average dose of the organs at risk to design an initial radiotherapy plan, and optimizing parameters of the objective function based on a particle swarm and genetic hybrid algorithm to obtain a parameter optimization result; and evaluating the parameter optimization result, updating the parameters of the objective function according to the evaluation result, and solving the objective function according to the parameter updating result to obtain the optimal radiotherapy plan. The invention realizes the function of automatically completing parameter optimization in the radiotherapy plan by an intelligent algorithm, and is convenient for improving the quality and efficiency of the plan.

Description

Automatic planning method and system based on dose prediction and parameter optimization
Technical Field
The invention relates to the technical field of radiotherapy treatment planning systems, in particular to an automatic planning method and system based on dose prediction and parameter optimization.
Background
The following two broad categories of current mainstream research directions for radiation therapy automatic planning techniques exist: (1) And simulating the design idea of a physicist by using a computer to realize automatic design of the plan. (2) Knowledge-based Planning (KBP) aims to improve the quality and efficiency of IMRT Planning by learning high-quality clinical Planning databases. The breast cancer automatic planning function in the RayStation of the commercial treatment planning system is to integrate a heuristic optimization algorithm into the TPS, the algorithm can effectively simulate all steps and decisions of a manual planning, an IMRT plan which can be clinically used is rapidly and automatically generated, and the quality and the efficiency of the planning are improved; a RapidPlanTM system developed by Waranan corporation trains a database by using a machine learning algorithm to obtain a DVH prediction model, and a physicist can predict a DVH distribution interval of a new plan by using the model and quickly obtain target function setting values such as volume dose relation, weight factors and the like.
Parameters of a target area and a endangered organ in the existing radiotherapy planning system are mainly adjusted manually, and the degree of intellectualization and automation is low. The optimization results of the currently used radiation therapy planning systems are largely determined by the experience and skill of the planning designer, and the optimization process takes a long time, and the physicist needs to spend a lot of time finding the most reasonable dose volume limit of the target region and the organs at risk and the weight corresponding to the limit. In studies of predicting three-dimensional dose distribution using a deep learning network, most scholars only achieve dose distribution prediction, and there are few studies on applying a predicted dose to generate a treatment plan.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, the invention aims to provide an automatic planning method based on dose prediction and parameter optimization, which solves the problems that a large amount of time is consumed when a radiotherapy plan is made, the quality of the plan highly depends on experience of a physicist and the like.
It is another object of the present invention to propose an automatic planning system based on dose prediction and parameter optimization.
In order to achieve the above object, the present invention provides an automatic planning method based on dose prediction and parameter optimization, including:
performing planned CT scanning on a patient to obtain a CT image, and obtaining organ delineation data according to the CT image;
inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction;
determining an objective function of an irradiation field distribution model based on the organ delineation data and the average dose of the organs at risk to design an initial radiotherapy plan, and optimizing parameters of the objective function based on a particle swarm and genetic hybrid algorithm to obtain a parameter optimization result;
and evaluating the parameter optimization result, updating the parameters of the objective function according to the evaluation result, and solving the objective function according to the parameter updating result to obtain the optimal radiotherapy plan.
The automatic planning method based on dose prediction and parameter optimization according to the embodiment of the present invention may also have the following additional technical features:
further, the inputting the CT image and the organ delineation data into the trained 3D VGG-U-Net network model to obtain the three-dimensional dose distribution prediction of the organ at risk includes: acquiring samples of CT images and organ delineation data, inputting the samples into a trained 3D VGG-U-Net network model for training, and obtaining the trained 3D VGG-U-Net network model; inputting the CT image and the organ delineation into the trained 3D VGG-U-Net network model, and performing relation construction operation among the CT image, the organ delineation data and dose distribution to obtain a relation construction result; wherein the relationship building operation comprises: convolution operation, batch standardization operation, deconvolution operation pooling operation, copying and splicing operation; and obtaining a three-dimensional dose distribution prediction of the organs at risk based on the relation construction result.
Further, the optimizing the parameters of the objective function based on the particle swarm and the genetic mixing algorithm to obtain a parameter optimization result includes: acquiring initial positions and initial speeds of all particles of the particle swarm, and calculating the fitness value of each particle according to a fitness function; comparing the fitness value of the current position of each particle with the fitness value of the current best position of the particle, and updating the fitness value of the current best position of the particle according to a first comparison result; comparing the fitness value of the current position of each particle with the fitness values of the best positions of all the current particles, and updating the fitness values of the best positions of all the current particles according to a second comparison result; if the adaptive values of the best positions of all the current particles do not meet the clinically prescribed dose target and the organ average dose target, adjusting the target area weight factor and the organ-at-risk average dose, and randomly assigning the values to certain particles; and updating the speed and the position of each particle according to a preset formula, enabling each particle and the randomly selected particles to be crossed at a random point according to a first preset probability, enabling each particle to be mutated at the random point according to a second preset probability, and taking the adaptive value of the best position of all the current particles as a global optimal solution.
Further, the method further comprises: and presetting the number and the angle of each irradiation field in the irradiation field distribution model, the initial irradiation field weight, the weight of a target function and the target area weight.
Further, the method further comprises: and optimizing the sub-field and the sub-field weight value of each irradiation field according to the parameter optimization of the objective function to obtain a sub-field optimization result and a sub-field weight value optimization result.
In order to achieve the above object, another aspect of the present invention provides an automatic planning system based on dose prediction and parameter optimization, comprising:
the data acquisition module is used for carrying out planned CT scanning on a patient to obtain a CT image and obtaining organ delineation data according to the CT image;
the dose prediction module is used for inputting the CT image and the organ delineation data into a trained 3DVGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction;
the parameter optimization module is used for determining an objective function of an irradiation field distribution model based on the organ delineation data and the average dose of the organs at risk so as to design an initial radiotherapy plan, and optimizing parameters of the objective function based on a particle swarm optimization and genetic hybrid algorithm to obtain a parameter optimization result;
and the plan output module is used for evaluating the parameter optimization result, updating the parameters of the objective function according to the evaluation result, and solving the objective function according to the parameter updating result to obtain the optimal radiotherapy plan.
According to the automatic planning method and system based on dose prediction and parameter optimization, disclosed by the embodiment of the invention, the parameter search range is limited by using a deep learning method, a physicist is replaced to complete automatic optimization of dosimetry parameters, and the efficiency and quality of planning a radiotherapy plan are improved.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flow diagram of an automated planning method based on dose prediction and parameter optimization according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a 3D VGG-U-Net network model structure according to an embodiment of the invention;
FIG. 3 is a flow chart of automatic optimization of parameters based on prediction of dose distribution according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an automatic planning system based on dose prediction and parameter optimization according to an embodiment of the present invention.
Detailed Description
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present invention will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, shall fall within the protection scope of the present invention.
An automatic planning method and system based on dose prediction and parameter optimization according to an embodiment of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a flow chart of an automated planning method based on dose prediction and parameter optimization in accordance with one embodiment of the present invention.
As shown in fig. 1, the method includes:
s1, performing planned CT scanning on a patient to obtain a CT image, and obtaining organ delineation data according to the CT image;
s2, inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction;
s3, determining an objective function of an irradiation field distribution model based on organ delineation data and the average dose of the organs at risk to design an initial radiotherapy plan, and optimizing parameters of the objective function based on a particle swarm and genetic hybrid algorithm to obtain a parameter optimization result;
and S4, evaluating the parameter optimization result, updating the parameters of the objective function according to the evaluation result, and solving the objective function according to the parameter updating result to obtain the optimal radiotherapy plan.
The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
It is understood that a radiotherapy Treatment Planning System (TPS) is a set of computer software systems that assist a physicist in designing a Treatment plan, and supports IMRT, VMAT, afterloading or proton Treatment, etc. as well as various computational mathematical models. Part of the TPS provides an Application Programming Interface (API) for the researcher, allowing the developer to read information such as patient data, dose matrix, and treatment plan through C # or python scripts, while allowing the radiation treatment plan to be modified and transmitted to the TPS system.
As an example, the method of the present invention may comprise the steps of:
(1) Planning a CT scan, obtaining the patient's outer contour, internal anatomy, and tumor volume from high quality CT images.
(2) The target area and the organs at risk are delineated, and the doctor or physicist delineates the anatomy of interest layer by layer from the CT image.
(3) Planning design, a physicist sets the number and the angle of irradiation fields, initial irradiation field weight, an objective function and a weight value thereof according to experience in a TPS system, and then a computer determines the sub-fields and the weight values thereof according to reverse optimization of the objective function. Plan design and other steps are performed alternately, a physicist can modify parameters such as an objective function weight value according to a plan evaluation result, then dose calculation and plan evaluation are performed again, and the process is repeated until the plan meets clinical requirements, and the process is a plan design flow and a core.
(4) Dose calculation: the dose calculation is performed using a convolution superposition algorithm or a monte carlo method.
(5) Evaluation and optimization: after planning, a physicist carries out planning evaluation according to evaluation functions such as dose volume histograms and the like, and if the evaluation functions do not meet clinical requirements, design and optimization are continued until an optimal treatment plan is found.
It can be understood that, the invention realizes the automatic optimization of the objective function parameters in the step (3), can effectively reduce the labor time, and reduce the difference of treatment plans formulated by different hospitals and different physicists, thereby improving the quality of the radiation treatment plan and further improving the curative effect and the survival quality of tumor patients.
Further, samples of the CT image and organ delineation data are obtained, and the samples are input into a trained 3D VGG-U-Net network model for training to obtain the trained 3D VGG-U-Net network model; inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model, and performing relation construction operation among the CT image, the organ delineation data and the dose distribution to obtain a relation construction result; and obtaining the three-dimensional dose distribution prediction of the organs at risk based on the relation construction result.
Specifically, multiple samples of CT images and organ delineation data are acquired for model training. A Convolutional Neural Network (CNN) is one of very representative structures in a deep learning Network, and is widely applied to the field of medical image processing. The invention provides a novel network structure 3D VGG-U-Net, which automatically learns the relation among CT images, organ delineation data and dose distribution through the network and realizes voxel-level three-dimensional dose distribution prediction. The network structure diagram is shown in fig. 2, where 3 × 3conv refers to a Convolution (Convolution) operation with a Convolution kernel size of 3 × 3; BN refers to the Batch Normalization (Batch Normalization) operation; relu refers to the activation function; 2 × 2deconv refers to a deconvolution operation with a convolution kernel size of 2 × 2; 2 x 2max Pooling refers to a Pooling (Pooling) operation with a core size of 2 x 2; copy & Connect refers to Copy and splice operations.
Further, the basic principle of Particle Swarm Optimization (PSO) is to simulate bird Swarm foraging behavior: consider that there is a food source in an area where a group of birds search for the food, they do not know the specific location of the food, but know how far away from the food source their current location is, and also know which bird is closest to the food source. At this time, the optimal search strategy is to search around the bird closest to the food. In the PSO algorithm, each potential solution of the optimization problem can be analogized to a bird, called a 'particle', and the particles update themselves according to the position of the current optimal particle and the position of the historical optimal particle, and search in the space until the space optimal solution is found.
The PSO algorithm has the following specific contents: first, N particles are initialized in a search space of D dimension, and an initial position and an initial velocity are randomly given to each particle, and both the position and the velocity are D dimension. And during the t-th iteration, determining the currently found optimal position pbest and the currently found optimal position gbest by calculating the fitness values of the particles. At t +1 iterations, each particle updates its velocity and position according to the following formula:
v t+1 ＝w·v t +c 1 ·r 1 ·(pbest t -x t )+c 2 ·r 2 ·(gbest t -x t ) (1-1)
x t+1 ＝x t +v t+1 (1-2)
in the formula, v t+1 Is the speed at t +1 iteration, v t Is the speed at the t-th iteration, x t+1 Is the position at t +1 th iteration, x t Is the position at the t-th iteration. w is an inertia factor, is a non-negative number, and adjusts the search range of the solution space; c. C 1 And c 2 To accelerateConstant, respectively regulating maximum step length of flight towards pbest and gbest directions, determining influence of individual experience and group experience of the particles on particle running tracks, and taking c in the invention 1 ＝c 2 ＝1.49445。r 1 And r 2 The random number is 0-1, and the randomness of particle flight is increased.
The rationale for Genetic Algorithms (GA) is to simulate the natural evolutionary process: each potential solution to the optimization problem can be analogized to a chromosome, with multiple chromosomes making up a population, and chromosomes evolving from one population to a new population through crossover and variation until a stop condition is met.
The PSO algorithm needs few parameters to be adjusted, is simple in structure and easy to realize, but lacks dynamic speed adjustment and is easy to fall into local optimum. The invention provides an improved particle swarm algorithm, namely a genetic particle swarm algorithm (GAPSO), which introduces cross variation operation in the GA algorithm into the PSO algorithm, reduces the probability of falling into local optimum and improves the algorithm performance. The improved particle swarm optimization algorithm of the invention has the following procedures:
(1) Initializing a particle swarm to obtain N particles, and endowing the N particles with random initial positions and initial speeds in a search space range;
(2) Calculating the fitness value of each particle according to the fitness function;
(3) For each particle, comparing the fitness value of the current position of the particle with the fitness value pbest of the best position at present, and if the fitness value of the current position is better, updating the pbest of the particle by using the current position;
(4) For each particle, the fitness value of the current position of the particle is compared with the fitness value gbest of the best position of all the particles so far, and if the fitness value of the current position is better, the gbest is updated by the current position;
(5) If the gbest does not meet the clinically prescribed dose target and the organ average dose target, the target area weight factor is increased by 10, the organ average dose at risk is decreased by 1.0Gy, and the value is randomly assigned to a particle;
(6) Updating the speed and position of each particle according to the formulas (1-1) and (1-2);
(7) Crossing each particle with randomly selected particles at a random point with a probability of 14%, and mutating each particle at the random point with a probability of 0.06%;
(8) And if the ending condition is not met, returning to the step 2, and if the ending condition is met, ending, wherein the gbest is the global optimal solution.
Furthermore, the invention inputs the CT image and organ delineation information into the trained 3D VGG-U-Net network, obtains three-dimensional dose distribution through prediction, obtains the average dose value of the organ according to the dose distribution, and the like, and can limit the search range of parameters during optimization by using the dose values, thereby improving the optimization speed. Then, outputting relevant data of a tumor target area and a dangerous organ by using an API (application programming interface) of the TPS, and automatically completing optimization of dose parameters such as preference degree, average dose and the like by using a hybrid algorithm of a particle swarm algorithm and a genetic algorithm to replace the process of manual adjustment of a physicist. And finally, inputting the optimized parameters into a planning system to generate a clinically acceptable radiotherapy plan. The optimization flow chart is shown in fig. 3.
According to the automatic planning method based on the dose prediction and the parameter optimization, disclosed by the embodiment of the invention, the parameter search range is limited by using a deep learning method, a physicist is replaced for completing the automatic optimization of the dosimetry parameters, and the efficiency and the quality of the radiation treatment plan formulation are improved.
In order to implement the above embodiment, as shown in fig. 4, an automatic planning system based on dose prediction and parameter optimization is further provided in this embodiment, where the system 10 includes: a data acquisition module 100, a dose prediction module 200, a parameter optimization module 300, and a plan output module 400.
The data acquisition module 100 is used for performing planned CT scanning on a patient to obtain a CT image and obtaining organ delineation data according to the CT image;
the dose prediction module 200 is used for inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction;
the parameter optimization module 300 is configured to determine an objective function of an irradiation field distribution model based on organ delineation information and an average dose of an organ at risk, to design an initial radiotherapy plan, and optimize parameters of the objective function based on a particle swarm optimization and a genetic hybrid algorithm to obtain a parameter optimization result;
and the plan output module 400 is configured to evaluate the parameter optimization result, update the parameter of the objective function according to the evaluation result, and solve the objective function according to the parameter update result to obtain the optimal radiotherapy plan.
Further, the dose prediction module 200 is further configured to:
acquiring samples of CT images and organ delineation data, inputting the samples into a trained 3D VGG-U-Net network model for training to obtain the trained 3D VGG-U-Net network model;
inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model, and performing a relation construction operation among the CT image, the organ delineation data and the dose distribution to obtain a relation construction result; wherein the relationship building operation comprises: convolution operation, batch standardization operation, deconvolution operation pooling operation, copying and splicing operation;
and obtaining the three-dimensional dose distribution prediction of the organs at risk based on the relation construction result.
Further, the parameter optimization module 300 is further configured to:
acquiring initial positions and initial speeds of all particles of the particle swarm, and calculating the fitness value of each particle according to the fitness function;
comparing the fitness value of the current position of each particle with the fitness value of the current best position of the particle, and updating the fitness value of the current best position of the particle according to a first comparison result; comparing the fitness value of the current position of each particle with the fitness values of the best positions of all the current particles, and updating the fitness values of the best positions of all the current particles according to a second comparison result;
if the adaptive values of the best positions of all the current particles do not meet the clinically prescribed dose target and the organ average dose target, adjusting the target area weight factor and the organ-at-risk average dose, and randomly assigning the values to certain particles;
and updating the speed and the position of each particle according to a preset formula, enabling each particle and the randomly selected particles to be crossed at a random point according to a first preset probability, enabling each particle to be mutated at the random point according to a second preset probability, and taking the adaptive value of the best position of all the current particles as a global optimal solution.
Further, the system 10 further includes an initial setting module, configured to preset the number and angle of each irradiation field in the irradiation field distribution model, an initial irradiation field weight, and a weight of the objective function and a target area weight.
Further, the system 10 further includes a sub-field optimization module, configured to optimize the sub-field and the sub-field weight value of each irradiation field according to parameter optimization of the objective function, so as to obtain a sub-field optimization result and a sub-field weight value optimization result.
According to the automatic planning system based on the dose prediction and the parameter optimization, disclosed by the embodiment of the invention, the parameter search range is limited by using a deep learning method, a physicist is replaced for completing the automatic optimization of the dosimetry parameters, and the efficiency and the quality of the radiation treatment plan formulation are improved.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.

Claims (10)

1. An automatic planning method based on dose prediction and parameter optimization is characterized by comprising the following steps:
performing planned CT scanning on a patient to obtain a CT image, and obtaining organ delineation data according to the CT image;
inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction;
determining an objective function of an irradiation field distribution model based on the organ delineation data and the average dose of the organs at risk to design an initial radiotherapy plan, and optimizing parameters of the objective function based on a particle swarm and genetic hybrid algorithm to obtain a parameter optimization result;
and evaluating the parameter optimization result, updating the parameters of the objective function according to the evaluation result, and solving the objective function according to the parameter updating result to obtain the optimal radiotherapy plan.
2. The method of claim 1, wherein the inputting the CT image and the organ delineation data into the trained 3D VGG-U-Net network model results in a three-dimensional dose distribution prediction of an organ at risk, comprising:
acquiring samples of CT images and organ delineation data, inputting the samples into a trained 3D VGG-U-Net network model for training, and obtaining the trained 3D VGG-U-Net network model;
inputting the CT image and the organ delineation data into the trained 3D VGG-U-Net network model, and performing a relation construction operation among the CT image, the organ delineation data and the dose distribution to obtain a relation construction result; wherein the relationship building operation comprises: convolution operation, batch standardization operation, deconvolution operation pooling operation, copying and splicing operation;
and obtaining a three-dimensional dose distribution prediction of the organs at risk based on the relation construction result.
3. The method of claim 1, wherein the optimizing parameters of the objective function based on the particle swarm optimization and the genetic mixture algorithm results in a parameter optimization result, comprising:
acquiring initial positions and initial speeds of all particles of the particle swarm, and calculating the fitness value of each particle according to the fitness function;
comparing the fitness value of the current position of each particle with the fitness value of the current best position of the particle, and updating the fitness value of the current best position of the particle according to a first comparison result; comparing the fitness value of the current position of each particle with the fitness values of the best positions of all the current particles, and updating the fitness values of the best positions of all the current particles according to a second comparison result;
if the adaptive values of the best positions of all the current particles do not meet the clinical prescription dose target and the organ average dose target, adjusting the target weight factor and the average dose of organs at risk, and randomly assigning the values to certain particles;
and updating the speed and the position of each particle according to a preset formula, enabling each particle to be crossed with randomly selected particles at a random point according to a first preset probability, enabling each particle to be mutated at the random point according to a second preset probability, and taking the adaptive value of the best position of all the current particles as a global optimal solution.
4. The method of claim 1, further comprising: and presetting the number and the angle of each irradiation field in the irradiation field distribution model, the initial irradiation field weight, the weight of the target function and the target area weight.
5. The method of claim 4, further comprising: and optimizing the sub-field and the sub-field weight value of each illumination field according to the parameter optimization of the objective function to obtain a sub-field optimization result and a sub-field weight value optimization result.
6. An automated planning system based on dose prediction and parameter optimization, comprising:
the data acquisition module is used for carrying out planned CT scanning on a patient to obtain a CT image and obtaining organ delineation data according to the CT image;
the dose prediction module is used for inputting the CT image and the organ delineation data into a trained 3D VGG-U-Net network model to obtain three-dimensional dose distribution prediction of the organs at risk, and obtaining the average dose of the organs at risk according to the three-dimensional dose distribution prediction;
the parameter optimization module is used for determining an objective function of an irradiation field distribution model based on the organ delineation data and the average dose of the organs at risk so as to design an initial radiotherapy plan, and optimizing parameters of the objective function based on a particle swarm optimization and genetic hybrid algorithm to obtain a parameter optimization result;
and the plan output module is used for evaluating the parameter optimization result, updating the parameters of the objective function according to the evaluation result, and solving the objective function according to the parameter updating result to obtain the optimal radiotherapy plan.
7. The system of claim 6, wherein the dose prediction module is further configured to:
acquiring samples of CT images and organ delineation data, inputting the samples into a trained 3D VGG-U-Net network model for training, and obtaining the trained 3D VGG-U-Net network model;
inputting the CT image and the organ delineation data into the trained 3D VGG-U-Net network model, and performing relation construction operation among the CT image, the organ delineation data and dose distribution to obtain a relation construction result; wherein the relationship building operation comprises: convolution operation, batch standardization operation, deconvolution operation pooling operation, copying and splicing operation;
and obtaining a three-dimensional dose distribution prediction of the organs at risk based on the relation construction result.
8. The system of claim 6, wherein the parameter optimization module is further configured to:
acquiring initial positions and initial speeds of all particles of the particle swarm, and calculating the fitness value of each particle according to the fitness function;
comparing the fitness value of the current position of each particle with the fitness value of the current best position of the particle, and updating the fitness value of the current best position of the particle according to a first comparison result; comparing the fitness value of the current position of each particle with the fitness values of the best positions of all the current particles, and updating the fitness values of the best positions of all the current particles according to a second comparison result;
if the adaptive values of the best positions of all the current particles do not meet the clinically prescribed dose target and the organ average dose target, adjusting the target area weight factor and the organ-at-risk average dose, and randomly assigning the values to certain particles;
and updating the speed and the position of each particle according to a preset formula, enabling each particle to be crossed with randomly selected particles at a random point according to a first preset probability, enabling each particle to be mutated at the random point according to a second preset probability, and taking the adaptive value of the best position of all the current particles as a global optimal solution.
9. The system of claim 6, further comprising an initial setting module for presetting the number and angles of each irradiation field, the initial irradiation field weight, and the weight of the objective function and the target area weight in the irradiation field distribution model.
10. The system of claim 9, further comprising a sub-field optimization module configured to optimize a sub-field and a sub-field weight value of each of the illumination fields according to the parameter optimization of the objective function, so as to obtain a sub-field optimization result and a sub-field weight value optimization result.
CN202211144237.8A 2022-09-20 2022-09-20 Automatic planning method and system based on dose prediction and parameter optimization Pending CN115620870A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211144237.8A CN115620870A (en) 2022-09-20 2022-09-20 Automatic planning method and system based on dose prediction and parameter optimization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211144237.8A CN115620870A (en) 2022-09-20 2022-09-20 Automatic planning method and system based on dose prediction and parameter optimization

Publications (1)

Publication Number Publication Date
CN115620870A true CN115620870A (en) 2023-01-17

Family

ID=84857852

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211144237.8A Pending CN115620870A (en) 2022-09-20 2022-09-20 Automatic planning method and system based on dose prediction and parameter optimization

Country Status (1)

Country Link
CN (1) CN115620870A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117132729A (en) * 2023-07-24 2023-11-28 清华大学 Multi-mode fine breast model design method, device, equipment and medium
CN117717723A (en) * 2024-02-08 2024-03-19 福建自贸试验区厦门片区Manteia数据科技有限公司 Portal information determining device, processor and electronic equipment

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117132729A (en) * 2023-07-24 2023-11-28 清华大学 Multi-mode fine breast model design method, device, equipment and medium
CN117717723A (en) * 2024-02-08 2024-03-19 福建自贸试验区厦门片区Manteia数据科技有限公司 Portal information determining device, processor and electronic equipment
CN117717723B (en) * 2024-02-08 2024-06-11 福建自贸试验区厦门片区Manteia数据科技有限公司 Portal information determining device, processor and electronic equipment

Similar Documents

Publication Publication Date Title
RU2684173C2 (en) System and method for automatic treatment planning
CN110211664B (en) System for automatically designing radiotherapy scheme based on machine learning
CN115620870A (en) Automatic planning method and system based on dose prediction and parameter optimization
US9507886B2 (en) Multi-objective radiation therapy optimization method
US20210339049A1 (en) Radiotherapy treatment plans using differentiable dose functions
CN107545137A (en) Tumour radiotherapy plan intelligent optimization method
CN110420396B (en) Radiotherapy plan optimization system, radiotherapy plan optimization device and storage medium
JP2007514499A (en) System and method for global optimization of treatment plans for external beam irradiation therapy
CN110124214A (en) Intensity modulated radiation therapy plan optimization method and application based on predicted dose distribution guidance
WO2020244172A1 (en) Plan implementation method and device based on predicted dose guidance and gaussian process optimization
Meyer et al. Application of influence diagrams to prostate intensity-modulated radiation therapy plan selection
US11738208B2 (en) Pareto optimal projection search (POPS) for automated radiation therapy treatment planning
US11282192B2 (en) Training deep learning engines for radiotherapy treatment planning
CN110706779B (en) Automatic generation method of accurate target function of radiotherapy plan
CN113795303A (en) Method and system for quality-aware continuous learning for radiation therapy planning
Nguyen et al. Advances in automated treatment planning
EP3642843A2 (en) Optimizing a treatment plan based on a score function depending on a quality function
CN115083567A (en) Rule-based automatic radiotherapy plan obtaining method and device
Wang et al. An integrated solution of deep reinforcement learning for automatic IMRT treatment planning in non-small-cell lung cancer
Bedford et al. Beam selection for stereotactic ablative radiotherapy using Cyberknife with multileaf collimation
CN117116421A (en) Method and device for determining radiotherapy plan
Lei et al. An approaching genetic algorithm for automatic beam angle selection in IMRT planning
CN113178242B (en) Automatic plan optimization system based on coupled generation countermeasure network
Deng et al. Neuro-dynamic programming for fractionated radiotherapy planning
Ghimire et al. Forecasting patient-specific dosimetric benefit from daily online adaptive radiotherapy for cervical cancer

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination