CN106682409B  Sampling method, radiotherapy plan optimization method and dose calculation method  Google Patents
Sampling method, radiotherapy plan optimization method and dose calculation method Download PDFInfo
 Publication number
 CN106682409B CN106682409B CN201611185113.9A CN201611185113A CN106682409B CN 106682409 B CN106682409 B CN 106682409B CN 201611185113 A CN201611185113 A CN 201611185113A CN 106682409 B CN106682409 B CN 106682409B
 Authority
 CN
 China
 Prior art keywords
 region
 sub
 sampling
 sampling points
 interest
 Prior art date
Links
 238000005070 sampling Methods 0.000 title claims abstract description 274
 238000005457 optimization Methods 0.000 title claims abstract description 42
 238000001959 radiotherapy Methods 0.000 title claims abstract description 19
 238000004364 calculation methods Methods 0.000 title abstract description 26
 238000004422 calculation algorithm Methods 0.000 claims abstract description 49
 238000007621 cluster analysis Methods 0.000 claims abstract description 29
 230000000875 corresponding Effects 0.000 claims description 24
 210000000056 organs Anatomy 0.000 abstract description 42
 206010028980 Neoplasm Diseases 0.000 description 9
 241000563949 Pentaphylacaceae Species 0.000 description 4
 230000004927 fusion Effects 0.000 description 4
 238000000034 methods Methods 0.000 description 4
 239000005020 polyethylene terephthalate Substances 0.000 description 4
 239000011159 matrix materials Substances 0.000 description 2
 239000010950 nickel Substances 0.000 description 2
 241001251094 Formica Species 0.000 description 1
 201000011510 cancer Diseases 0.000 description 1
 210000004027 cells Anatomy 0.000 description 1
 230000004907 flux Effects 0.000 description 1
 239000002245 particles Substances 0.000 description 1
 230000000717 retained Effects 0.000 description 1
 238000002922 simulated annealing Methods 0.000 description 1
 210000001519 tissues Anatomy 0.000 description 1
 210000004881 tumor cells Anatomy 0.000 description 1
Classifications

 A—HUMAN NECESSITIES
 A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
 A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
 A61N5/00—Radiation therapy
 A61N5/10—Xray therapy; Gammaray therapy; Particleirradiation therapy
 A61N5/103—Treatment planning systems
 A61N5/1031—Treatment planning systems using a specific method of dose optimization

 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
 G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
 G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computeraided diagnosis, e.g. based on medical expert systems
Abstract
Description
Technical Field
The invention relates to the field of radiotherapy, in particular to a clusteringbased sampling method in a radiotherapy plan.
Background
Radiotherapy is a method of treating malignant tumors using radiation, such as α, β, gamma rays generated by radioisotopes, and xrays, electron beams, proton beams, and other particle beams generated by various types of xray treatment machines or accelerators.
Due to the high beam energy, normal cells are affected while tumor cells are killed. In order to minimize damage to normal tissues, radiation treatment plans need to be developed. In the radiation treatment plan, a doctor needs to give a prescription and a treatment plan, and a physicist draws the organ and the tumor position, and Target areas such as total tumor Volume (GTV), Clinical Target Volume (CTV) and Planned Treatment Volume (PTV) according to the doctor prescription, and makes and optimizes the radiation treatment plan.
In order to determine the quality of the radiation treatment plan, a dose calculation algorithm (e.g., monte carlo algorithm, volume calculation method, pencil beam algorithm, etc.) is generally used to simulate the distribution of the dose received by various parts of the human body according to the radiation treatment plan.
Both optimization of the radiation treatment plan and dose calculation need to be based on the image, e.g. optimization of the radiation treatment plan and dose calculation based on CT images. The CT images generally have higher resolution and a larger number of pixels, which results in too long computation time and too large memory space if all the pixels participate in the optimization of the radiation treatment plan and the dose calculation.
In order to solve the problems, the CT image can be subjected to downsampling, and sampling points represent all pixel points to participate in optimization of a radiation treatment plan and calculation of dose distribution, so that the selection of the sampling points directly influences a calculation result.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a method for sampling images in radiation treatment plan optimization, which comprises the following steps: the method comprises the steps of obtaining an image to be sampled, wherein the image to be sampled comprises at least one region of interest, obtaining the upper limit of the total number of sampling points and the number of the sampling points contained in each subregion in the region of interest, dividing the region of interest into a plurality of subregions according to the upper limit of the total number of the sampling points and the number of the sampling points contained in each subregion in the region of interest, and performing downsampling on each subregion by using a cluster analysis algorithm to obtain the sampling points of the region.
Optionally, dividing the region of interest into a plurality of subregions according to the upper limit of the total number of the sampling points and the number of the sampling points included in each subregion in the region of interest includes: dividing the region of interest into a plurality of subregions according to a preset rule, calculating the sum of the number of sampling points contained in each subregion, comparing the sum of the number of sampling points contained in each subregion with the upper limit of the total number of the sampling points, if the sum of the number of sampling points contained in each subregion is greater than the upper limit of the total number of the sampling points, expanding the plurality of subregions, calculating the sum of the number of sampling points contained in each expanded subregion, comparing the sum of the number of sampling points contained in each expanded subregion with the upper limit of the total number of the sampling points until the sum of the number of sampling points contained in each expanded subregion is less than the upper limit of the total number of the sampling points, wherein the subregion in the last cycle is the final subregion, if, and reducing the plurality of subareas and calculating the sum of the number of sampling points contained in each reduced subarea, comparing the sum of the number of sampling points contained in each reduced subarea with the upper limit of the total number of the sampling points until the sum of the number of sampling points contained in each reduced subarea is greater than or equal to the upper limit of the total number of the sampling points, wherein each subarea in the current cycle is the final subarea, and if the sum of the number of sampling points contained in each subarea is equal to the upper limit of the total number of the sampling points, each subarea is the final subarea.
Optionally, the number of sampling points included in the subregion is related to at least one of the type of the region of interest, the weight, and the distance between the region of interest and the target region.
Optionally, the downsampling each subregion by using the cluster analysis algorithm includes: sampling each subregion according to the number of sampling points contained in each subregion in the region of interest to obtain initial sampling points, calculating Euler distances from each pixel point to the initial sampling points in each subregion, classifying each pixel point into a cluster corresponding to the initial sampling point with the minimum Euler distance, and iteratively executing the following steps in each subregion until the downsampling of each subregion is completed: and calculating the center of each cluster as a new sampling point, calculating the Euler distance between the new sampling point and the last sampling point, comparing the Euler distance with a threshold value, taking the new sampling point as the sampling point of the subregion if the Euler distance does not exceed the threshold value, calculating the Euler distance from each pixel point to the new sampling point if the Euler distance exceeds the threshold value, and classifying each pixel point into the cluster corresponding to the new sampling point with the minimum Euler distance.
Optionally, before downsampling each subregion by using a cluster analysis algorithm, it is determined whether the number of pixel points belonging to the corresponding region of interest in the subregion exceeds the number of corresponding sampling points, if not, all the pixel points belonging to the corresponding region of interest in the subregion are taken as the sampling points, and if so, downsampling the pixel points belonging to the corresponding region of interest in the subregion by using the cluster analysis algorithm.
Optionally, before performing downsampling on each subregion by using a cluster analysis algorithm, it is determined whether pixel points not belonging to the corresponding region of interest are included in the subregion, if so, the pixel points belonging to the corresponding region of interest in the subregion are downsampled by using the cluster analysis algorithm, if not, the subregion is uniformly divided into grids according to the number of sampling points, and the central point of each grid is used as the sampling point.
According to another embodiment, the present invention further provides a method for sampling images in radiation treatment plan optimization, comprising: the method comprises the steps of obtaining an image to be sampled, wherein the image to be sampled comprises an interested region, obtaining the upper limit of the total number of sampling points, determining the number of the sampling points of each interested region, and performing downsampling on the interested region by utilizing a cluster analysis algorithm to obtain the sampling points of the interested region.
According to another embodiment, the present invention also provides a radiation therapy plan optimization method, including: the method comprises the steps of obtaining an image of a patient and sketching interested areas, obtaining a dosage target of each interested area, sampling each interested area by using any one of the sampling methods, establishing an optimization model and optimizing so that the dosage of sampling points meets the dosage target.
According to another embodiment, the present invention also provides a radiation therapy plan optimization system, comprising: the system comprises an acquisition unit, a delineation unit, an input unit, a sampling unit and an optimization unit, wherein the acquisition unit is used for acquiring a patient image, the delineation unit is used for delineating interested areas, the input unit is used for setting dosage targets of the interested areas, the sampling unit is used for sampling the interested areas by using the sampling method, and the optimization unit is used for establishing an optimization model and optimizing the optimization model so that the dosage of sampling points meets the dosage targets.
According to another embodiment, the present invention also provides a dose calculation method including: acquiring a patient image, delineating interested regions, sampling the interested regions by using the sampling method, and calculating the dose distribution of the interested regions based on the sampling points.
Compared with the prior art, the method for sampling the image in the optimization of the radiation treatment plan clusters the pixels based on the characteristics of the pixels, and selects one pixel point from various types as the sampling point, so that the shape characteristics of each organ or target area can be kept while the number of the sampling points is reduced, and the effectiveness of the radiation treatment plan and the accuracy of dose distribution calculation are improved;
by dividing the region of interest into a plurality of subregions, pixels are clustered in each subregion based on the characteristics of the pixels, thereby reducing the amount of computation and improving the computation speed.
The sampling method is applied to optimization of the radiation treatment plan, which is beneficial to improving the optimization speed and effectiveness of the radiation treatment plan and enables the calculation result of the dose distribution to be more consistent with the actual dose distribution.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a flowchart of a first method for image sampling in radiation treatment plan optimization according to an embodiment of the present invention;
FIG. 2 is a flowchart of a second method for image sampling in radiation treatment plan optimization according to an embodiment of the present invention;
FIG. 3 is a flow chart of a radiation treatment plan optimization method provided by an embodiment of the present invention;
fig. 4 is a flowchart of a dose calculation method according to an embodiment of the present invention.
Detailed Description
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 derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In radiation therapy, it is desirable to deliver as high a dose of radiation as possible to the target tumor volume, minimizing damage to the organs. It is therefore desirable to optimize the radiation treatment plan and calculate the dose distribution in the patient to which the radiation treatment plan corresponds. Both the optimization of the radiation treatment plan and the dose distribution are calculated based on the sampling points and are therefore influenced by the distribution of the sampling points. In order to better control the dose distribution within each organ or target volume, the embodiments of the present invention propose a new sampling method.
Fig. 1 is a flowchart of a method for sampling images in radiation treatment plan optimization according to an embodiment of the present invention. Referring to fig. 1, the sampling method 100 provided by the present invention includes:
and step S101, acquiring an image to be sampled.
Acquiring an image to be sampled, wherein the image to be sampled comprises at least one region of interest, and the region of interest comprises an organ in which a doctor is interested and a tumor target area.
The patient image may be a CT image, a PET image, an MR image, or a fusion image, etc., but is not limited thereto.
And step S102, acquiring the upper limit of the total number of the sampling points and determining the number of the sampling points of each region of interest.
The upper limit N of the total number of sampling points limits the total number of sampling points of the at least one region of interest, which can be set in advance.
The number of the sampling points of each interested area is related to the type, weight and size of the interested area, for example, the interested area can be a tumor target area and can also be an organ, the weight of the target area is not less than that of the organ, and the weight of the organ at risk is more than that of the common organ; under the condition that the weight is kept unchanged, the larger the interested region is, the more the number of the sampling points is, therefore, the number of the sampling points of each organ and the target region is determined according to the type, the weight and the size of each interested region, and the condition that the sum of the number of the sampling points of each interested region does not exceed the upper limit N of the total number of the sampling points is met.
And S103, performing downsampling on the region of interest by using a cluster analysis algorithm to obtain sampling points of the region of interest.
And sampling each organ and the target area according to the number of the sampling points of each interested area determined in the step. The number of sampling points of the ith organ or target area is recorded as N_{i}The sampling algorithm is used for sampling each organ and the target area, and the sampling algorithm is not limited herein. For the sake of simplicity, the process of downsampling the region of interest by using the cluster analysis algorithm in this embodiment is described here by taking a random sampling algorithm as an example, but the scope of the invention is not limited thereto.
Suppose that N is randomly selected from all pixel points in the ith organ_{i}Each sampling point, calculating the Euler distance from each pixel point to each sampling point, and grouping each pixel point to the cluster corresponding to the sampling point with the minimum Euler distance, thereby forming N_{i}And (4) recalculating the center of each cluster as a new sampling point, and comparing the new sampling point with the last sampling point to judge whether the convergence condition is met.
The center of the cluster can be calculated by using a kmeans algorithm, for example, the coordinates of each pixel point in a certain cluster are respectively (x)_{1},y_{1}),(x_{2},y_{2}),…,(x_{n},y_{n}) And n is the number of pixel points in the cluster, the center coordinate of the cluster isThe center of the cluster may also be calculated by using a kcenter algorithm, for example, for a certain cluster, the sum of euler distances from each pixel point to other pixel points is calculated, and the point with the minimum sum of the euler distances is used as the center of the cluster.
And judging whether the Euler distance between the new sampling point and the last determined sampling point does not exceed a threshold value, if not, finishing the circulation, taking the new sampling point as the sampling point of the region of interest, and if not, continuing the circulation until the Euler distance between the new sampling point and the last sampling point does not exceed the threshold value. The threshold value can here be set empirically beforehand.
In this embodiment, a clustering analysis algorithm is applied to sampling of an image, pixels are clustered based on distances, the closer the distance, the more similar the pixels are in representing the shape of an area of interest, and a pixel point can be selected from each class as a sampling point, so that the shape characteristics of the area of interest can be retained while the number of the sampling points is reduced.
Fig. 2 is a flowchart of a method for sampling images in radiation treatment plan optimization according to an embodiment of the present invention. Referring to fig. 2, the sampling method 200 provided by the present invention includes:
s201, acquiring an image to be sampled.
Acquiring an image to be sampled, wherein the image to be sampled comprises at least one region of interest, and the region of interest comprises an organ in which a doctor is interested and a tumor target area.
The patient image may be a CT image, a PET image, an MR image, or a fusion image, etc., but is not limited thereto.
In this embodiment, only the pixel points belonging to the region of interest are sampled, so that the region of interest can be represented by a matrix. The pixel points in the region of interest are represented by "1" and the pixels outside the region of interest are represented by "0", so that each region of interest can be represented by a matrix. Subsequent processing may only process areas covered by "1".
Step S202, acquiring the upper limit of the total number of the sampling points and the number of the sampling points contained in each subarea in the region of interest.
The upper limit N of the total number of sampling points limits the total number of sampling points in the region of interest, which can be set by the physician.
The number of sampling points in each subregion in the region of interest determines the number of sampling points in the region of interest, and can be set by a doctor. The number of sampling points in the subregions may be different in different organs or target regions, depending on at least one of the type and weight of the region of interest, and the distance between the region of interest and the target region, for example, if the type of the region of interest includes an organ and a target region, the number of sampling points in the subregions in the target region is not less than the number of sampling points in the subregions in the organ; the weight of the target area is not less than the weight of the organ, the weight of the organ at risk is higher than the weight of the common organ, the number of sampling points of the subarea in the target area is not less than the number of sampling points of the subarea in the organ at risk, and the number of sampling points of the subarea in the organ at risk is greater than the number of sampling points of the subarea in the common organ; the closer to the target area, the greater the number of sample points for subareas within the region of interest. Considering at least one factor, the number of sampling points of each subregion in the target area can be set to be 24, for example 4, the number of sampling points of the subregion in the organ at risk is 23, for example 2, and the number of sampling points of the subregions in other organs is 1.
And step S203, dividing the region of interest into a plurality of subregions according to the upper limit of the total number of the sampling points and the number of the sampling points contained in each subregion in the region of interest.
And dividing each organ of interest and the target area into a plurality of subareas according to a preset rule. For the convenience of calculation, all organs and target regions may be divided into a plurality of subregions by using the same rule, for example, each organ and target region may be divided by using an equidistant grid, and the side length of the grid is a multiple of the pixel pitch. Each organ and target area are divided into equidistant grids, for example with 10 times the pixel pitch. The preset rule may be set in advance, and the content of the rule is not limited herein.
And calculating the sum of the number of the sampling points contained in each subregion according to the number of the sampling points contained in each subregion acquired in the step S202. For example, each subregion of the target region includes 4 sampling points, each subregion of the organ at risk includes 2 sampling points, each subregion of the other organs includes 1 sampling point, and the number of sampling points included in all subregions is added to obtain the sum of the number of sampling points.
And comparing the sum of the number of the sampling points obtained by calculation with the upper limit of the total number of the sampling points, and judging whether the size of the subregion needs to be adjusted according to the comparison result.
If the sum of the number of the sampling points contained in each subregion is greater than the upper limit of the total number of the sampling points, the plurality of subregions are expanded, the sum of the number of the sampling points contained in each expanded subregion is calculated, the sum of the number of the sampling points contained in each expanded subregion is compared with the upper limit of the total number of the sampling points until the sum of the number of the sampling points contained in each expanded subregion is less than the upper limit of the total number of the sampling points, and the subregion in the last cycle is the final subregion. The subregions may be enlarged in predetermined steps, for example by increasing the length of the grid side by one pixel pitch at a time.
If the sum of the number of the sampling points contained in each subarea is smaller than the upper limit of the total number of the sampling points, reducing the subareas and calculating the sum of the number of the sampling points contained in each reduced subarea, comparing the sum of the number of the sampling points contained in each reduced subarea with the upper limit of the total number of the sampling points until the sum of the number of the sampling points contained in each reduced subarea is larger than or equal to the upper limit of the total number of the sampling points, and each subarea in the current cycle is the final subarea. The subregions may be reduced in predetermined steps, for example by reducing the grid edge length by the length of one pixel pitch at a time.
And if the sum of the number of the sampling points contained in each subregion is equal to the upper limit of the total number of the sampling points, each current subregion is the final subregion.
The at least one region of interest may be partitioned by an iterative loop process as described above. However, the present invention is not limited thereto, and it is within the protection scope of the present invention to divide the region of interest by using other manners according to the upper limit of the total number of the sampling points and the number of the sampling points contained in each subregion in the region of interest.
In this embodiment, the image to be sampled may be uniformly divided, or each organ and the target area may be individually divided, which is not limited herein.
And S204, performing downsampling on each subregion by using a cluster analysis algorithm to obtain sampling points of the region of interest.
Taking a certain subregion as an example, sampling the subregion according to the number of corresponding sampling points to obtain initial sampling points. The sampling algorithm is not limited herein. For the sake of simplicity, the process of downsampling the region of interest by using the cluster analysis algorithm in this embodiment is described here by taking a random sampling algorithm as an example, but the scope of the invention is not limited thereto.
Calculating the Euler distance from each pixel point in the subregion to the initial sampling point, classifying each pixel point to a cluster corresponding to the initial sampling point with the minimum Euler distance, recalculating the center of each cluster as a new sampling point, calculating the Euler distance between the new sampling point and the last sampling point, and comparing the Euler distance with a threshold value to judge whether the convergence condition is met.
The center of the cluster can be calculated by using a kmeans algorithm, for example, the coordinates of each pixel point in a certain cluster are respectively (x)_{1},y_{1}),(x_{2},y_{2}),…,(x_{n},y_{n}) And n is the number of pixel points in the cluster, the center coordinate of the cluster isThe center of the cluster may also be calculated by using a kcenter algorithm, for example, for a certain cluster, the sum of euler distances from each pixel point to other pixel points is calculated, and the point with the minimum sum of the euler distances is used as the center of the cluster.
Judging whether the Euler distance between a new sampling point and the last determined sampling point does not exceed a threshold value, if not, taking the new sampling point as the sampling point of the subregion, if so, calculating the Euler distance between a pixel point in the subregion and the new sampling point, classifying each pixel point into a cluster corresponding to the new sampling point with the minimum Euler distance, calculating the central point of each cluster again, taking the central point as the new sampling point, comparing the central point with the original sampling point until the Euler distance between the new sampling point and the sampling point in the last cycle is smaller than the threshold value, and taking the new sampling point as the sampling point of the subregion. The threshold value in this embodiment may be set empirically in advance.
The above steps are performed iteratively within each subregion, thereby completing the sampling of each organ and target region.
In the embodiment, each organ or target area is divided into a plurality of subareas, and downsampling is performed in the subareas by using a cluster analysis algorithm, so that not only can sampling points for reserving the shape characteristics of the region of interest be obtained, but also the calculation amount can be reduced, and the calculation speed can be improved.
After the region of interest is divided into a plurality of subregions, the number of pixel points contained in the subregions is limited. In the above embodiment, before the cluster analysis algorithm is used to perform downsampling on the subregion, it may be determined whether the number of the pixel points belonging to the region of interest in the subregion exceeds the number of the corresponding sampling points, if not, all the pixel points belonging to the region of interest in the subregion are taken as the sampling points, and if so, the cluster analysis algorithm is used to perform downsampling on the pixel points belonging to the region of interest in the subregion.
After dividing the region of interest into a plurality of subregions, the subregions may contain pixel points that do not belong to the organ (pixel points represented by "0") near the edge of the region of interest. In the above embodiment, before the subregion is downsampled by using the cluster analysis algorithm, it may be determined whether a pixel point not belonging to the region of interest is included in the subregion, if so, the pixel point not belonging to the region of interest in the subregion is downsampled by using the cluster analysis algorithm, if not, the subregion may be uniformly divided into a corresponding number of grids according to the number of sampling points, and the central point of the grid is used as the sampling point, so as to further reduce the amount of computation, improve the computation speed, and save time.
In the above embodiment, before the subregion is downsampled by using the cluster analysis algorithm, it may be determined whether the subregion includes a pixel that does not belong to the region of interest, if yes, judging whether the number of the pixel points belonging to the interested region in the subregion exceeds the number of the corresponding sampling points, if not, all the pixel points belonging to the interested region in the region are taken as sampling points, if the number of the pixels in the subregion exceeds the threshold value, the pixel points belonging to the interesting region in the subregion are downsampled by utilizing a cluster analysis algorithm, if not, the subregion can be uniformly divided into a corresponding number of grids according to the number of the sampling points, the central point of the grid is taken as the sampling point, therefore, the calculation amount is further reduced, the calculation speed is improved, the time is saved, and other sampling algorithms can be adopted to directly sample the subregion.
The method for sampling the region of interest based on the cluster analysis algorithm provided by the above embodiment can be applied to a radiation therapy plan optimization method. Fig. 3 is a flowchart of a radiation treatment plan optimization method according to an embodiment of the present invention. Referring to fig. 3, the radiation treatment plan optimization method 300 includes:
s301, acquiring a patient image and delineating the region of interest.
An image of the patient is loaded and a region of interest within the image is delineated. In this embodiment, the patient image may be loaded into a radiation therapy planning system, and a region of interest, including an organ of interest to the physician and a tumor target, is delineated on an interface of the radiation therapy planning system. The sketching can be manually sketched by a doctor or automatically sketched, or the manual sketching and the automatic sketching are combined.
The patient image may be a CT image, a PET image, an MR image, or a fusion image, etc., but is not limited thereto.
S302, obtaining the dose target of each interested area.
The dose targets for each region of interest may be set by the physician. Prior to radiation treatment planning, the physician sets the prescribed dose for each region of interest, such as the upper organatrisk dose limit and the lower and upper tumor target volume dose limits. In planning the radiation treatment, the radiation treatment planning system receives the prescribed dose as a dose target.
And S303, downsampling each region of interest.
The method in fig. 1 or fig. 2 in the above embodiment may be adopted to perform downsampling on each region of interest based on a cluster analysis algorithm.
S304, establishing an optimization model and optimizing so that the dose of the sampling points meets the dose target.
In an embodiment of the present invention, the objective function f (d) of the sampling pointbased radiation therapy plan optimization can be expressed by the following formula:
wherein d is_{v}Current dose, u, for sampling point v_{v}Upper dose limit for sampling point v,/_{v}Lower dose limit for sampling point v.
And optimizing the objective function by using an optimization algorithm (such as a simulated annealing algorithm, a gradient algorithm, an ant colony algorithm and the like), so that the dose of each sampling point meets the dose constraint and an optimized radiation flux map is obtained.
Correspondingly, an embodiment of the present invention further provides a radiation therapy plan optimization system, including:
an acquisition unit for acquiring an image of a patient,
a delineation unit for delineating the region of interest,
an input unit for setting a dose target for each region of interest,
a sampling unit for sampling the regions of interest by using the method of the above embodiment in FIG. 1 or FIG. 2,
and the optimization unit is used for establishing an optimization model and optimizing the optimization model so that the dosage of the sampling point meets the dosage target.
The method and system for optimizing the radiation therapy plan in the embodiment retain the shape characteristics of each organ and the target region based on the sampling points, so that the obtained radiation therapy plan is high in effectiveness.
The method for sampling the region of interest based on the cluster analysis algorithm provided by the above embodiment can be applied to dose calculation. Fig. 4 is a flowchart of a dose calculation method according to an embodiment of the present invention. Referring to fig. 4, the dose calculation method 400 includes:
s401, acquiring a patient image and delineating the region of interest.
An image of the patient is loaded and a region of interest within the image is delineated, including the organ of interest of the physician and the tumor target area. The sketching can be manually sketched by a doctor or automatically sketched, or the manual sketching and the automatic sketching are combined.
The image may be a CT image, a PET image, an MR image, a fusion image, or the like, but is not limited thereto.
S402, downsampling each region of interest.
The method in fig. 1 or fig. 2 in the above embodiment may be adopted to perform downsampling on each region of interest based on a cluster analysis algorithm.
And S403, calculating the dose distribution of each region of interest based on the sampling points.
The dose distribution in each organ and target volume is calculated based on the sampling points to obtain the dose distribution in the patient's body corresponding to the radiation treatment plan, which can be provided to a physician for determining the quality of the radiation treatment plan. There are many algorithms for dose calculation, such as monte carlo algorithm, volume calculation method, pencil beam algorithm, etc., and there is no limitation here.
Because the characteristics of organs and target areas are reserved by the sampling points obtained based on the clustering analysis algorithm, the dose distribution obtained based on the sampling point calculation is more consistent with the actual dose distribution, and the accuracy of the dose distribution calculation is improved.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.
Claims (6)
Priority Applications (1)
Application Number  Priority Date  Filing Date  Title 

CN201611185113.9A CN106682409B (en)  20161220  20161220  Sampling method, radiotherapy plan optimization method and dose calculation method 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

CN201611185113.9A CN106682409B (en)  20161220  20161220  Sampling method, radiotherapy plan optimization method and dose calculation method 
Publications (2)
Publication Number  Publication Date 

CN106682409A CN106682409A (en)  20170517 
CN106682409B true CN106682409B (en)  20200331 
Family
ID=58869804
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

CN201611185113.9A CN106682409B (en)  20161220  20161220  Sampling method, radiotherapy plan optimization method and dose calculation method 
Country Status (1)
Country  Link 

CN (1)  CN106682409B (en) 
Families Citing this family (2)
Publication number  Priority date  Publication date  Assignee  Title 

CN109125952A (en) *  20180718  20190104  中北大学  Convolution based on nuclear model is superimposed dose calculation methodology 
CN109034136A (en) *  20180906  20181218  湖北亿咖通科技有限公司  Image processing method, device, picture pickup device and storage medium 
Citations (3)
Publication number  Priority date  Publication date  Assignee  Title 

CN104548372A (en) *  20150107  20150429  上海联影医疗科技有限公司  Radiotherapy planning method and device, radiotherapy dose determining method and device and radiotherapy quality guaranteeing method and device 
CN105404789A (en) *  20151228  20160316  上海联影医疗科技有限公司  Selection method and selection system for dose control points, and radiotherapy planning system 
CN105561485A (en) *  20151231  20160511  上海联影医疗科技有限公司  Radiotherapy treatment planning optimization method and device 
Family Cites Families (1)
Publication number  Priority date  Publication date  Assignee  Title 

US9409039B2 (en) *  20130521  20160809  Varian Medical Systems International Ag  Systems and methods for automatic creation of dose prediction models and therapy treatment plans as a cloud service 

2016
 20161220 CN CN201611185113.9A patent/CN106682409B/en active IP Right Grant
Patent Citations (3)
Publication number  Priority date  Publication date  Assignee  Title 

CN104548372A (en) *  20150107  20150429  上海联影医疗科技有限公司  Radiotherapy planning method and device, radiotherapy dose determining method and device and radiotherapy quality guaranteeing method and device 
CN105404789A (en) *  20151228  20160316  上海联影医疗科技有限公司  Selection method and selection system for dose control points, and radiotherapy planning system 
CN105561485A (en) *  20151231  20160511  上海联影医疗科技有限公司  Radiotherapy treatment planning optimization method and device 
NonPatent Citations (1)
Title 

不平衡数据的降采样方法研究;林舒杨等;《计算机研究与发展》;20111231;第48卷;4753页 * 
Also Published As
Publication number  Publication date 

CN106682409A (en)  20170517 
Similar Documents
Publication  Publication Date  Title 

CN109074500B (en)  System and method for segmenting medical images of the same patient  
JP6688536B2 (en)  Systems and methods for learning models of radiation therapy treatment planning and predicting radiation therapy dose distributions  
US20190362522A1 (en)  Neural network for generating synthetic medical images  
CN104812442B (en)  The system, apparatus and method of quality assurance for radiotherapy  
JP6620116B2 (en)  System and method for automated treatment planning  
US10773101B2 (en)  System and method for estimating and manipulating estimated radiation dose  
US9555265B2 (en)  Intensitymodulated ion therapy  
Korsholm et al.  A criterion for the reliable use of MRIonly radiotherapy  
Unkelbach et al.  Robust radiotherapy planning  
Pötter et al.  Present status and future of highprecision image guided adaptive brachytherapy for cervix carcinoma  
US9123097B2 (en)  System and method for automatic generation of initial radiation treatment plans  
US10369381B2 (en)  System and method for robust intensitymodulated proton therapy planning  
AU2016338923B2 (en)  PseudoCT generation from MR data using a feature regression model  
Hou et al.  Deformable planning CT to cone‐beam CT image registration in head‐and‐neck cancer  
Elstrøm et al.  Daily kV conebeam CT and deformable image registration as a method for studying dosimetric consequences of anatomic changes in adaptive IMRT of head and neck cancer  
EP1778074B1 (en)  Weighted surfacetosurface mapping  
US6200255B1 (en)  Prostate implant planning engine for radiotherapy  
Hellebust et al.  Dosimetric impact of interobserver variability in MRIbased delineation for cervical cancer brachytherapy  
RU2629235C2 (en)  Dose calculation at beam segments level and motion tracking in time for adaptive planning of treatment  
US9327139B2 (en)  Radiation therapy treatment plan improvement through use of knowledge base  
Bauer et al.  Integration and evaluation of automated Monte Carlo simulations in the clinical practice of scanned proton and carbon ion beam therapy  
US8121252B2 (en)  Use of planning atlas in radiation therapy  
JP6138804B2 (en)  Dose calculation using inhomogeneitycompensated superposition for radiation therapy  
CN109906470A (en)  Use the image segmentation of neural network method  
CN102306239B (en)  Method for evaluating and optimizing radiotherapy dose based on cone beam CT (Computer Tomography) image CT value correction technology 
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  
CP01  Change in the name or title of a patent holder 
Address after: 201807 No. 2258 Chengbei Road, Jiading Industrial Zone, Jiading District, Shanghai. Patentee after: Shanghai Lianying Medical Technology Co., Ltd Address before: 201807 No. 2258 Chengbei Road, Jiading Industrial Zone, Jiading District, Shanghai. Patentee before: SHANGHAI UNITED IMAGING HEALTHCARE Co.,Ltd. 