CN107715314A - Radiotherapy system and method based on deep learning - Google Patents
Radiotherapy system and method based on deep learning Download PDFInfo
- Publication number
- CN107715314A CN107715314A CN201711158702.2A CN201711158702A CN107715314A CN 107715314 A CN107715314 A CN 107715314A CN 201711158702 A CN201711158702 A CN 201711158702A CN 107715314 A CN107715314 A CN 107715314A
- Authority
- CN
- China
- Prior art keywords
- radiotherapy
- patient
- module
- convolutional neural
- neural networks
- 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.)
- Granted
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
-
- 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
- A61N2005/1041—Treatment planning systems using a library of previously administered radiation treatment applied to other patients
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Radiation-Therapy Devices (AREA)
Abstract
The present invention discloses a kind of radiotherapy system and method based on deep learning, including medical cases database module, image data processing module, radiotherapy data processing module, depth convolutional neural networks module, Radiation treatment plans generation module.The present invention can combine the image data and part radiation therapy data of patient, the radiotherapy treatment planning of the existing patient of study, so as to directly generate radiotherapy treatment planning scheme for new patient, greatly save the time that doctor debugs radiotherapy treatment planning repeatedly, reduce labor intensity and the professional standards requirement to operating personnel, radiotherapy treatment planning preferably is provided for patient, and then improves the effect of radiation therapy.
Description
Technical field
The invention belongs to tumour radiotherapy technical field, is related to a kind of side for automatically generating tumour radiotherapy plan
Method, specifically a kind of radiotherapy system and method based on deep learning.
Background technology
Deep learning is by establishing multilayer neural network, imitating the study mechanism of human brain, be issued in the training of data
Can automatic data processing, auxiliary or substitution people complete the task of high intensity human-computer interaction.The application neck of deep learning at present
In image recognition processing, voice technology, unmanned etc., do not applied on radiotherapy also mainly in domain.
Radiotherapy treatment planning is according to the clinical characters of patient, three-dimensional/four-dimensional medical image Tomography and not
With radiation equipmen etc. formulate patient detailed therapeutic scheme and (including target area, jeopardize organ, dose of radiation, ray energy
With size, incident angle and time etc.).The radiotherapy planning of early stage is a kind of " planning by hand " process, with computer application in
Planning process, dose distributions computation work is completed by computer, and undertaker is responsible for selecting free parameter.This be one repeatedly
Trial and error process, to be gone on always before suitable scheme is found.The limitation of this semi-hand mode is also apparent.
More it is rational plan thought should be：The parameter configuration optimal from the dose prescription set off in search for the treatment of expert's formulation, and on
State way and be only able to find a feasible program, without optimization process.Optimization method used is broadly divided into radiotherapy
Linear programming technique and Nonlinear Programming Method.In linear programming technique, conventional has simplex method.And in Nonlinear Programming Method, often
There are square optimization, gradient projection method, Constrained simulated annealing etc..(Chinese biomedical engineering journal volume 21 the 4th
The optimum algorithm of multi-layer neural network of phase radiotherapy treatment planning)
A kind of Chinese patent " optimization intensity modulated radiation therapy meters based on image feature and intelligent regression model of CN102184330A-
The method drawn " provides a kind of method of the optimization intensity modulated radiation therapy plan based on image feature and intelligent regression model, builds first
The database of vertical expert's level intensity modulated radiation therapy plan, the record of each intensity modulated radiation therapy case is preserved in the lane database；Then root
According between the existing medical image feature and intensity modulated radiation therapy planning optimization parameter of the lane database of expert's level intensity modulated radiation therapy plan
Related information, pass through Supervised machine learning establish the intensity modulated radiation therapy planning optimization parameter based on medical image feature intelligence
Regression model；Graphical analysis is carried out to the medical image of specific intensity modulated radiation therapy object, and carries out feature extraction；By characteristics of image
Intelligent regression model is inputted, exports intensity modulated radiation therapy planning optimization parameter；Set according to result above on treatment planning systems (TPS)
Set the tone strong radiotherapy planning Optimal Parameters, run TPS planning optimization function, obtain the optimum treatment plan of radiotherapy.
The radiotherapy planning software that Hospitals at Present uses needs radiotherapy expert to be co-operated with medical physicist, adjusts ginseng to repair repeatedly
Just, it is necessary to expend a large amount of manpowers, higher is required to the specialty of operating personnel.
A kind of Chinese patent " optimization intensity modulated radiation therapy meters based on image feature and intelligent regression model of CN102184330A-
The method drawn " is not directed to depth learning technology crucial herein, it is impossible to and it is intelligent to data with existing storehouse progress maximum to utilize, it is required
Manpower is more；This method can only optimize the parameter of radiotherapy planning, it is impossible to realize the design of complete radiotherapy planning.
The content of the invention
In place of overcoming the shortcomings of the prior art, the present invention provides a kind of radiotherapy system based on deep learning
System and method, automatically generate radiotherapy treatment planning, to can quickly realize the individual character of tumor patient using depth learning technology
Chemoradiotherapy scheme, improve the survival rate of tumor patient.
The purpose of the present invention can be achieved through the following technical solutions：
Radiotherapy system based on deep learning, including：Medical cases database module, image data processing module,
Radiotherapy data processing module, depth convolutional neural networks module, Radiation treatment plans generation module；
The medical cases database module：Store the medical image data and radiation therapy data of patient, above-mentioned storage
Training set of the information as depth convolutional neural networks；
The image data processing module：The medical image data of patient is read, and performs corresponding pretreatment operation, so
Pretreated image information is sent to depth convolutional neural networks module afterwards；
The radiotherapy data processing module：The part radiotherapy data of patient are read, and perform corresponding vectorization operation, so
The numerical information after processing is sent to depth convolutional neural networks module afterwards；
The depth convolutional neural networks module：Receive from image data processing module and radiotherapy data processing module
Data information, and neutral net is inputted, by neural computing, export remaining radiotherapy data parameters；
The Radiation treatment plans generation module：Receive image data processing module, radiotherapy data processing module and depth convolution
The data information of neural network module, TPS systems are then inputted, calculate the dosage distribution of patient, and showed to doctor
Final radiotherapy treatment planning scheme.
Described medical cases database module, includes two parts content：
When the image data of patient, including CT images data and MRI image data；
Second, the radiotherapy data of patient, include the parameter of radiotherapy equipment, the position of therapeutic bed and angle, patient's tumour
Position coordinates, size, target area and jeopardize organ dosage limitation, therapeutic modality, radioactive source energy, launched field quantity, ray class
Type, the angle and weight of ray.Input of a portion parameter as neutral net, another part parameter is as neutral net
Label.
Described image data processing module carries out 3-dimensional resampling, a patient after processing to patient's image of input
Number of sections is N, pixel size H*W.
Described radiotherapy data processing module is converted to the part radiotherapy data that doctor inputs the vector of one S dimension, makees
For radiotherapy characteristic parameter.
Described deep neural network training when using supervised mode of learning, by asking a regularization
The minimum value of cost function and provide optimal solution, in the training process, adjusted using error back propagation and gradient descent method
Network parameter, the effect of whole convolutional neural networks can be represented with a functional relation：Y=h (X)
Wherein X represents the patient's image data and part radiotherapy data of input, and Y represents remaining radiotherapy data parameters, h letters
Number represents depth convolutional neural networks.
Described Radiation treatment plans generation module can receive image data processing module, radiotherapy data processing module and depth automatically
The data information of convolutional neural networks module is spent, is then inputted TPS systems, the dosage distribution of patient is calculated, so as to obtain
A set of radiotherapy planning scheme.
A kind of radiation therapy method based on deep learning, comprises the following steps：
Step 1：Medical cases database is established, enough training datas are provided for depth convolutional neural networks；
Step 2：Depth convolutional neural networks are trained, using supervised mode of learning, by seeking a cost function most
Small value and provide optimal solution；
Step 3：The medical files of patient are selected, import radiotherapy system, and inputs the part related to the patient and puts
Data are treated, image data processing module can read patient's medical image information automatically, and image is pre-processed, radiotherapy data
Processing module can read the patient part radiotherapy parameter information of input automatically, obtain corresponding radiotherapy characteristic parameter；
Step 4：Depth convolutional neural networks module receives patient's image and radiotherapy characteristic parameter, passes through what is had built up
Deep neural network exports the remaining radiotherapy characteristic of the patient；
Step 5：Radiation treatment plans generation module receives image data and whole radiotherapy parameters, and optimization is calculated patient and shone
The dosage distributed intelligence of emitter official, so as to obtain a set of radiotherapy planning scheme.
Step 6：Doctor audits the radiotherapy planning scheme, if examination ＆ verification passes through, implements the radiotherapy planning scheme, and by the patient
Radiotherapy planning scheme adds medical cases database；If examination ＆ verification is not by adjusting the part radiotherapy parameter of input, recalculating.
Compared with the prior art, beneficial effects of the present invention are embodied in：
The present invention realizes a kind of radiotherapy system using depth learning technology, is that deep learning is applied into radiotherapy
Once breakthrough trial in plan design, not only make use of the image information of patient, have also combined the part radiotherapy ginseng of patient
Number information, the feature of study is more diversified, and by learning the radiotherapy treatment planning of existing patient, it is new that can allow computer
Patient directly generates radiotherapy planning parameter, greatlys save the time for debugging radiation treatment parameters repeatedly, reduces labor intensity
And the professional standards requirement to operating personnel, while improve the quality of Chemical Examination Material in Hospital treatment plan；
The image data processing module and radiotherapy data processing module that the present invention develops, patient's image information and part are put
Treat parameter information to combine, realize the intelligent utilization of maximum to data with existing；
The deep neural network module that the present invention develops, make use of depth learning technology, image that can be from patient and portion
Radiotherapy parameter information learning is divided to change traditional radiation therapy to the other radiotherapy parameters how set in radiotherapy treatment planning
Plan the mode that design process traditional Chinese medical science green hand moves tuning parameter, greatly improve the efficiency of radiotherapy doctor；
The medical cases database module that the present invention develops, the continuous expansion of database is realized, often diagnoses a patient,
Patient information and treatment plan will be automatically added in database, data set can constantly increase, and often increase certain amount disease
People's information, depth convolutional neural networks can be trained again, so improve constantly software performance, so as to generate more preferable treatment
Plan, form a benign cycle.
Brief description of the drawings
For the ease of it will be appreciated by those skilled in the art that the present invention is further illustrated below in conjunction with the accompanying drawings.
Fig. 1 is the DFD of the radiotherapy system of the invention based on deep learning；
Fig. 2 is the flow chart of the radiation therapy method of the invention based on deep learning.
Embodiment
As shown in figure 1, a kind of radiotherapy system based on deep learning, including：Medical cases database module, image
Data processing module, radiotherapy data processing module, depth convolutional neural networks module, Radiation treatment plans generation module；
Medical cases database module：Store the medical image data and radiation therapy data of patient, the letter of above-mentioned storage
Cease the training set as depth convolutional neural networks；
Image data processing module：The medical image data of patient is read, and performs corresponding pretreatment operation, then will
Pretreated image information is sent to depth convolutional neural networks module；
Radiotherapy data processing module：The part radiotherapy data of patient are read, and perform corresponding vectorization operation, then will
Numerical information after processing is sent to depth convolutional neural networks module；
Depth convolutional neural networks module：Receive the data from image data processing module and radiotherapy data processing module
Data, and neutral net is inputted, by neural computing, export remaining radiotherapy data parameters；
Radiation treatment plans generation module：Receive image data processing module, radiotherapy data processing module and depth convolutional Neural
The data information of mixed-media network modules mixed-media, TPS systems are then inputted, calculate the dosage distribution of patient, and showed finally to doctor
Radiotherapy treatment planning scheme.
Described medical cases database module, includes two parts content：
First, the image data of patient, the CT images number of sections of each patient is N (such as 128, or other are by designer
Defined numerical value), picture element matrix size is H*W (such as 256*256, or other numerical value as defined in designer)；
Second, the parameter of the radiotherapy data, wherein radiotherapy equipment of patient, the position of therapeutic bed and angle, patient's tumour
Position coordinates, size, and target area and jeopardize the dosage of organ and limit input as neutral net, its numerical information is pressed
The vector of a S (such as 64, or other numerical value as defined in designer) dimension is stored as according to unified form；Remaining radiotherapy parameter is such as
Radioactive source energy, launched field quantity, ray type, the angle and weight of ray, label of the therapeutic modality as neutral net, by it
Numerical information is stored as the vector of a T dimension (such as 96, or other numerical value as defined in designer) according to unified form.
Described image data processing module can carry out 3-dimensional resampling, a disease after processing to patient's image of input
People's number of sections is N (such as 128, or other numerical value as defined in designer), pixel size be H*W (such as 256*256, or
Other numerical value as defined in designer),
The machine parameter that described radiotherapy data processing module can input doctor, treats bed position and angle, supplement
The position of patient's tumour and size, and given target area and jeopardize organ dose limit require information be converted to a S (such as
64, or other numerical value as defined in designer) vector of dimension.
Described depth convolutional neural networks module includes a 3D convolutional neural networks, be divided into X layers (such as 10 layers,
Or other numerical value as defined in designer), be below 10 layers of neutral net preferred embodiment：
● first layer is made up of three substratums：
1) 3D convolutional layers
2) ReLu active coatings
3) 3D maximums pond layer
● the same first layer of the second Rotating fields
● third layer is made up of two substratums：
1) 3D convolutional layers
2) ReLu active coatings
● the same first layer of four-layer structure
● the same third layer of layer 5 structure
● the same first layer of layer 6 structure
● the same first layer of layer 7 structure
● the 8th layer is made up of three substratums：
1) full articulamentum
2) ReLu active coatings
3) drop layers
● the 9th layer is made up of three substratums：
1) fused layer
2) full articulamentum
3) drop layers
● the tenth layer is made up of a full articulamentum, and it is T to set neuromere to count, and as output layer, exports remaining radiotherapy
Parameter.
Described depth convolutional neural networks training when using supervised mode of learning, by asking a regularization
Cost function minimum value and provide optimal solution, in the training process, using error back propagation and gradient descent method
Network parameter is adjusted, the effect of whole convolutional neural networks can be represented with a functional relation：Y=h (X)
Wherein X represents the patient's image data and part radiotherapy data of input, and Y represents remaining radiotherapy data parameters, h letters
Number represents depth convolutional neural networks.
Described Radiation treatment plans generation module can receive image data processing module, radiotherapy data processing module and depth automatically
The data information of convolutional neural networks module is spent, is then inputted TPS systems, the dosage distribution of patient is calculated, so as to obtain
A set of radiotherapy planning scheme.
A kind of radiation therapy method based on deep learning of the present invention, comprises the following steps：
Step 1：Medical cases database is established, enough training datas are provided for depth convolutional neural networks, due to god
There is unified requirement to data format through network, we will be carried out to the ill information of patient and the treatment information of doctor at unified
Manage and stored, database mainly includes following two parts content：
First, the image data of patient, the CT images number of sections of each patient is N (such as 128, or other are by designer
Defined numerical value), picture element matrix size is H*W (such as 256*256, or other numerical value as defined in designer)；
Second, the parameter of the radiotherapy data, wherein radiotherapy equipment of patient, the position of therapeutic bed and angle, patient's tumour
Position coordinates, size, and target area and jeopardize the dosage of organ and limit input as neutral net, its numerical information is pressed
The vector of a S (such as 64, or other numerical value as defined in designer) dimension is stored as according to unified form；Remaining radiotherapy parameter is such as
Radioactive source energy, launched field quantity, ray type, the angle and weight of ray, label of the therapeutic modality as neutral net, by it
Numerical information is stored as the vector of a T dimension (such as 96, or other numerical value as defined in designer) according to unified form；
Step 2：Depth convolutional neural networks are trained, using supervised mode of learning, by seeking a cost function most
Small value and provide optimal solution, in order to prevent over-fitting, the cost function that described cost function has been regularization, formula is as follows：
Wherein X represents the patient's image and part radiotherapy information of input, and θ then joins for each layer of depth convolutional neural networks
Number, in the training process, according to error back propagation, network parameter is adjusted using gradient descent method, its formula is as follows：
When neutral net restrains or during up to highest iterations, training stops, and now, the training of neutral net is completed, whole
The effect of individual convolutional neural networks can be represented with a functional relation：Y=h (X), wherein X represent patient's image of input
Data and part radiotherapy data, Y represent remaining radiotherapy data parameters, h function representation depth convolutional neural networks；
Step 3：The medical files of a patient are selected, such as DICOM file, import the software, and input radiotherapy apparatus
The parameter of device, the position of therapeutic bed and angle, the position coordinates of patient's tumour, size, target area and the dosage limitation for jeopardizing organ.
Image data processing module can read medical image information automatically, and carry out 3-dimensional resampling to image, be unified for N*H*W (such as
128*256*256, or other numerical value as defined in designer) specified format, radiotherapy data processing module can read defeated automatically
The radiotherapy data message entered, and its numerical information is stored as a S (such as 64, or other are by designer according to unified form
Defined numerical value) dimension vector；
Step 4：Depth convolutional neural networks module receives patient's image and radiotherapy characteristic parameter, passes through what is had built up
Deep neural network exports the remaining radiotherapy characteristic of the patient；
Step 5：Radiation treatment plans generation module receives image data and whole radiotherapy parameters, and optimization is calculated patient and shone
The dosage distributed intelligence of emitter official, so as to obtain a set of radiotherapy planning scheme.
Step 6：Doctor audits the radiotherapy planning scheme, if examination ＆ verification passes through, implements the radiotherapy planning scheme, and by the patient
Radiotherapy planning scheme adds medical cases database；If examination ＆ verification is not by adjusting the part radiotherapy parameter of input, recalculating.
The present invention can combine the image information and radiotherapy parameter information of patient, learn the radiotherapy treatment planning of existing patient,
So as to directly generate radiotherapy parameter for new patient, the time that doctor debugs radiotherapy parameter repeatedly is greatlyd save, reduces labor
Fatigue resistance and the professional standards requirement to operating personnel, while the quality of Chemical Examination Material in Hospital treatment plan is improved, carried for patient
Higher living guarantee is supplied.
Present invention disclosed above preferred embodiment is only intended to help and illustrates the present invention.Preferred embodiment is not detailed
All details are described, it is only described embodiment also not limit the invention.Obviously, according to the content of this specification,
It can make many modifications and variations.This specification is chosen and specifically describes these embodiments, is to preferably explain the present invention
Principle and practical application so that skilled artisan can be best understood by and utilize the present invention.The present invention is only
Limited by claims and its four corner and equivalent.
Claims (8)
1. the radiotherapy system based on deep learning, it is characterised in that including：Medical cases database module, image data
Processing module, radiotherapy data processing module, depth convolutional neural networks module, Radiation treatment plans generation module；
The medical cases database module：Store the medical image data and radiation therapy data of patient, the letter of above-mentioned storage
Cease the training set as depth convolutional neural networks；
The image data processing module：The medical image data of patient is read, and performs corresponding pretreatment operation, then will
Pretreated image information is sent to depth convolutional neural networks module；
The radiotherapy data processing module：The part radiotherapy data of patient are read, and perform corresponding vectorization operation, then will
Numerical information after processing is sent to depth convolutional neural networks module；
The depth convolutional neural networks module：Receive the data from image data processing module and radiotherapy data processing module
Data, and neutral net is inputted, by neural computing, export remaining radiotherapy data parameters；
The Radiation treatment plans generation module：Receive image data processing module, radiotherapy data processing module and depth convolutional Neural
The data information of mixed-media network modules mixed-media, TPS systems are then inputted, calculate the dosage distribution of patient, and showed finally to doctor
Radiotherapy treatment planning scheme.
2. the radiotherapy system based on deep learning as claimed in claim 1, it is characterised in that described medical cases number
According to library module, two parts content is included：
When the image data of patient, including CT images data and MRI image data；
Second, the radiotherapy data of patient, include the parameter of radiotherapy equipment, the position of therapeutic bed and angle, the position of patient's tumour
Coordinate, size, target area and the dosage limitation for jeopardizing organ are put, therapeutic modality, radioactive source energy, launched field quantity, ray type, is penetrated
The angle and weight of line；Input of a portion parameter as neutral net, mark of another part parameter as neutral net
Label.
3. the radiotherapy system based on deep learning as claimed in claim 1, it is characterised in that：At described image data
Manage module and 3-dimensional resampling is carried out to patient's image of input, a patient slices quantity after processing is N, pixel size H*
W。
4. the radiotherapy system based on deep learning as claimed in claim 1, it is characterised in that：At described radiotherapy data
Reason module is converted to the part radiotherapy data that doctor inputs the vector of one T dimension, as radiotherapy characteristic parameter.
5. the radiotherapy system based on deep learning as claimed in claim 1, it is characterised in that：Described depth convolution god
Include a 3D convolutional neural networks through mixed-media network modules mixed-media, there is a many levels structure, including convolutional layer, pond layer, local articulamentum,
Full articulamentum, fused layer, reshape layers, dropout layers, active coating.
6. the radiotherapy system based on deep learning as claimed in claim 1, it is characterised in that：Described depth nerve net
Network in training using supervised mode of learning, by asking the minimum value of cost function of a regularization to provide most
Excellent solution, in the training process, network parameter, whole convolutional neural networks are adjusted using error back propagation and gradient descent method
Effect can be represented with a functional relation：
Y=h (X)
Wherein X represents the patient's image data and part radiotherapy data of input, and Y represents remaining radiotherapy data parameters, h function tables
Show depth convolutional neural networks.
7. the radiotherapy system based on deep learning as claimed in claim 1, it is characterised in that：Described Radiation treatment plans life
The data of image data processing module, radiotherapy data processing module and depth convolutional neural networks module can be received automatically into module
Data, TPS systems are then inputted, the dosage distribution of patient are calculated, so as to obtain a set of radiotherapy planning scheme.
8. the radiation therapy method based on deep learning, it is characterised in that：Comprise the following steps：
Step 1：Medical cases database is established, enough training datas are provided for depth convolutional neural networks；
Step 2：Depth convolutional neural networks are trained, using supervised mode of learning, by the minimum value for seeking a cost function
And provide optimal solution；
Step 3：The medical files of patient are selected, import radiotherapy system, and input the part radiotherapy number related to the patient
According to image data processing module can read patient's medical image information automatically, and image is pre-processed, radiotherapy data processing
Module can read the patient part radiotherapy parameter information of input automatically, obtain corresponding radiotherapy characteristic parameter；
Step 4：Depth convolutional neural networks module receives patient's image and radiotherapy characteristic parameter, passes through the depth having had built up
Neutral net exports the remaining radiotherapy characteristic of the patient；
Step 5：Radiation treatment plans generation module receives image data and whole radiotherapy parameters, and patient's exposure device is calculated in optimization
The dosage distributed intelligence of official, so as to obtain a set of radiotherapy planning scheme.
Step 6：Doctor audits the radiotherapy planning scheme, if examination ＆ verification passes through, implements the radiotherapy planning scheme, and by patient's radiotherapy
Plans add medical cases database；If examination ＆ verification is not by adjusting the part radiotherapy parameter of input, recalculating.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711158702.2A CN107715314B (en) | 2017-11-20 | 2017-11-20 | Deep learning-based radiotherapy system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711158702.2A CN107715314B (en) | 2017-11-20 | 2017-11-20 | Deep learning-based radiotherapy system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107715314A true CN107715314A (en) | 2018-02-23 |
| CN107715314B CN107715314B (en) | 2020-04-14 |
Family
ID=61217652
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711158702.2A Active CN107715314B (en) | 2017-11-20 | 2017-11-20 | Deep learning-based radiotherapy system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107715314B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108717866A (en) * | 2018-04-03 | 2018-10-30 | 陈辛元 | A kind of method, apparatus, equipment and the storage medium of the distribution of prediction radiotherapy planning dosage |
| CN109166613A (en) * | 2018-08-20 | 2019-01-08 | 北京东方瑞云科技有限公司 | Radiotherapy treatment planning assessment system and method based on machine learning |
| CN109621228A (en) * | 2018-12-12 | 2019-04-16 | 上海联影医疗科技有限公司 | The calculating unit and storage medium of radiological dose |
| CN109919210A (en) * | 2019-02-26 | 2019-06-21 | 华南理工大学 | A kind of heart sound semisupervised classification method based on depth convolutional network |
| CN110706779A (en) * | 2019-10-10 | 2020-01-17 | 福建省肿瘤医院（福建省肿瘤研究所、福建省癌症防治中心） | Automatic generation method of accurate target function of radiotherapy plan |
| CN110960803A (en) * | 2018-09-28 | 2020-04-07 | 瓦里安医疗系统国际股份公司 | Method and system for adaptive radiation treatment planning using a deep learning engine |
| CN111145901A (en) * | 2019-12-04 | 2020-05-12 | 深圳大学 | Deep venous thrombosis thrombolytic curative effect prediction method and system, storage medium and terminal |
| CN112037885A (en) * | 2020-09-07 | 2020-12-04 | 平安科技（深圳）有限公司 | Dose prediction method, device, computer equipment and storage medium in radiotherapy planning |
| CN112546463A (en) * | 2021-02-25 | 2021-03-26 | 四川大学 | Radiotherapy dose automatic prediction method based on deep neural network |
| CN113226459A (en) * | 2018-12-24 | 2021-08-06 | 皇家飞利浦有限公司 | Automated detection for monitoring pulmonary condition of a chest patient receiving external beam radiation therapy |
| WO2022088341A1 (en) * | 2020-10-30 | 2022-05-05 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Radiation therapy control method and apparatus |
| US11682485B2 (en) | 2018-09-28 | 2023-06-20 | Siemens Healthineers International Ag | Methods and systems for adaptive radiotherapy treatment planning using deep learning engines |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170193175A1 (en) * | 2015-12-30 | 2017-07-06 | Case Western Reserve University | Prediction of recurrence of non-small cell lung cancer |
| CN107072624A (en) * | 2014-06-18 | 2017-08-18 | Elekta股份有限公司 | system and method for automatic treatment plan |
-
2017
- 2017-11-20 CN CN201711158702.2A patent/CN107715314B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107072624A (en) * | 2014-06-18 | 2017-08-18 | Elekta股份有限公司 | system and method for automatic treatment plan |
| US20170193175A1 (en) * | 2015-12-30 | 2017-07-06 | Case Western Reserve University | Prediction of recurrence of non-small cell lung cancer |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108717866A (en) * | 2018-04-03 | 2018-10-30 | 陈辛元 | A kind of method, apparatus, equipment and the storage medium of the distribution of prediction radiotherapy planning dosage |
| CN108717866B (en) * | 2018-04-03 | 2022-10-11 | 中国医学科学院肿瘤医院 | Method, device, equipment and storage medium for predicting radiotherapy plan dose distribution |
| CN109166613A (en) * | 2018-08-20 | 2019-01-08 | 北京东方瑞云科技有限公司 | Radiotherapy treatment planning assessment system and method based on machine learning |
| CN110960803A (en) * | 2018-09-28 | 2020-04-07 | 瓦里安医疗系统国际股份公司 | Method and system for adaptive radiation treatment planning using a deep learning engine |
| US11682485B2 (en) | 2018-09-28 | 2023-06-20 | Siemens Healthineers International Ag | Methods and systems for adaptive radiotherapy treatment planning using deep learning engines |
| CN110960803B (en) * | 2018-09-28 | 2022-02-25 | 瓦里安医疗系统国际股份公司 | Computer system for performing adaptive radiation therapy planning |
| US11167152B2 (en) | 2018-12-12 | 2021-11-09 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for generating a dose distribution |
| CN109621228A (en) * | 2018-12-12 | 2019-04-16 | 上海联影医疗科技有限公司 | The calculating unit and storage medium of radiological dose |
| US11813480B2 (en) | 2018-12-12 | 2023-11-14 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for generating a dose distribution |
| CN113226459A (en) * | 2018-12-24 | 2021-08-06 | 皇家飞利浦有限公司 | Automated detection for monitoring pulmonary condition of a chest patient receiving external beam radiation therapy |
| CN109919210A (en) * | 2019-02-26 | 2019-06-21 | 华南理工大学 | A kind of heart sound semisupervised classification method based on depth convolutional network |
| CN110706779A (en) * | 2019-10-10 | 2020-01-17 | 福建省肿瘤医院（福建省肿瘤研究所、福建省癌症防治中心） | Automatic generation method of accurate target function of radiotherapy plan |
| CN110706779B (en) * | 2019-10-10 | 2022-06-14 | 福建省肿瘤医院（福建省肿瘤研究所、福建省癌症防治中心） | Automatic generation method of accurate target function of radiotherapy plan |
| CN111145901B (en) * | 2019-12-04 | 2021-02-09 | 深圳大学 | Deep venous thrombosis thrombolytic curative effect prediction method and system, storage medium and terminal |
| CN111145901A (en) * | 2019-12-04 | 2020-05-12 | 深圳大学 | Deep venous thrombosis thrombolytic curative effect prediction method and system, storage medium and terminal |
| CN112037885B (en) * | 2020-09-07 | 2022-05-17 | 平安科技（深圳）有限公司 | Dose prediction method, device, computer equipment and storage medium in radiotherapy planning |
| CN112037885A (en) * | 2020-09-07 | 2020-12-04 | 平安科技（深圳）有限公司 | Dose prediction method, device, computer equipment and storage medium in radiotherapy planning |
| WO2022088341A1 (en) * | 2020-10-30 | 2022-05-05 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Radiation therapy control method and apparatus |
| CN112546463A (en) * | 2021-02-25 | 2021-03-26 | 四川大学 | Radiotherapy dose automatic prediction method based on deep neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107715314B (en) | 2020-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107715314A (en) | Radiotherapy system and method based on deep learning | |
| Chen et al. | A feasibility study on an automated method to generate patient‐specific dose distributions for radiotherapy using deep learning | |
| Liu et al. | A cascade 3D U‐Net for dose prediction in radiotherapy | |
| CN106029170B (en) | The establishment of increment treatment plan | |
| Ma et al. | Incorporating dosimetric features into the prediction of 3D VMAT dose distributions using deep convolutional neural network | |
| Chanyavanich et al. | Knowledge‐based IMRT treatment planning for prostate cancer | |
| Petrovic et al. | Knowledge-light adaptation approaches in case-based reasoning for radiotherapy treatment planning | |
| CN110197709A (en) | A kind of 3-dimensional dose prediction technique based on deep learning Yu priori plan | |
| CN102184330A (en) | Image feature and smart regressive model-based method for optimizing intensity modulated radiotherapy plan | |
| CN104815392B (en) | A kind of interactive radiotherapy treatment planning system optimization system | |
| Alber et al. | Intensity modulated radiotherapy treatment planning by use of a barrier-penalty multiplier method | |
| Kajikawa et al. | Automated prediction of dosimetric eligibility of patients with prostate cancer undergoing intensity-modulated radiation therapy using a convolutional neural network | |
| US11786204B2 (en) | Automatically-registered patient fixation device images | |
| CN109801696A (en) | A kind of cloud radiotherapy planning method, equipment, storage medium and the system of artificial intelligence | |
| Morén et al. | Mathematical optimization of high dose-rate brachytherapy—derivation of a linear penalty model from a dose-volume model | |
| CN105561485B (en) | The optimization device of radiotherapy treatment planning | |
| Ma et al. | A feasibility study on deep learning‐based individualized 3D dose distribution prediction | |
| Fan et al. | Independent verification of brachytherapy treatment plan by using deep learning inference modeling | |
| CN110021399A (en) | A kind of automatic design method of radiotherapy treatment planning | |
| Kawamura et al. | Revolutionizing radiation therapy: the role of AI in clinical practice | |
| Wang et al. | An integrated solution of deep reinforcement learning for automatic IMRT treatment planning in non-small-cell lung cancer | |
| Marungo et al. | Creating a data science platform for developing complication risk models for personalized treatment planning in radiation oncology | |
| Kallis et al. | Automated treatment planning framework for brachytherapy of cervical cancer using 3D dose predictions | |
| CN110465004A (en) | A kind of generation method of cloud radiotherapy treatment planning system and radiotherapy treatment planning | |
| CN104252580B (en) | A kind of method that radiotherapy planning transplanting is realized on Pinnacle treatment planning systems |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |