Radiotherapy system, data processing method and storage medium

1. A radiotherapy system, comprising:

a calibration database which is accessed to a clinical database through a preset database interface to acquire in real time or periodically a plurality of sample data generated in a clinical therapy process;

a training module which is connected to the calibration database through a preset interface to perform deep learning on any one of the plurality of sample data stored in the calibration database to obtain a therapy algorithm model corresponding to any one of the therapy stages, and sends the therapy algorithm model to a data processing module;

the data processing module configured to process the detection data received in the clinical therapy process according to the therapy algorithm model to generate preliminary therapy data;

wherein each sample data comprises a set of detection data and a set of therapy data corresponding to the set of the detection data and applied to the clinical therapy process and the therapy data is verified in the clinical therapy process;

the clinical therapy process comprises a therapy plan generation stage, a pre-therapy positioning stage and an in-therapy monitoring stage;

in the therapy plan generation stage, the detection data comprises medical image data, the therapy data comprises a therapy plan, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the therapy plan generation stage to obtain a therapy plan formulation model;

in the pre-therapy positioning stage, the detection data comprises medical image data acquired before the clinical therapy process and in the clinical therapy process, the therapy data comprises a positioning offset of a patient, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the pre-therapy positioning stage to obtain a pre-therapy positioning model;

in the in-therapy monitoring stage, the detection data comprises a patient displacement, a tumor monitoring image, a remaining radiation dosage, the therapy data comprises a tumor displacement before and after therapy, a dosage error, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the in-therapy monitoring stage to obtain a monitoring model;

the calibration database classifies and stores the detection data and the therapy data generated in the therapy plan generation stage, the pre-therapy positioning stage and the in-therapy monitoring stage;

wherein:

the training module is further configured to receive a feedback level of a specified sample data, and the feedback level is positively related to a therapeutic effect of therapy data in the specified sample data; and

the training module is further configured to adjust a weight value of the specified sample data according to the feedback level of the specified sample data, wherein the adjusted weight value of the specified sample data is positively related to the feedback level of the specified sample data, and perform deep learning on sample data with the adjusted weight value in the calibration database; and

wherein:

the data processing module comprises an image registration sub-module, a delineation sub-module, a therapy plan formulation sub-module, a positioning sub-module and a monitoring sub-module;

the training module performs deep learning on a plurality of medical image data and registration images generated at an image registration stage to obtain a registration model; and

the image registration sub-module processes medical image data acquired in the clinical therapy process according to the registration model and outputs a preliminary registration image;

the training module performs deep learning on the plurality of registration images and delineation result images generated at a delineation stage to obtain a delineation model; and the delineation sub-module processes registration images acquired in the clinical therapy process according to the delineation model and outputs a preliminary delineation result image;

the positioning sub-module processes the medical image data acquired at the image registration stage and the medical image data acquired at the pre-therapy positioning stage during the clinical therapy process according to the pre-therapy positioning model and outputs a preliminary positioning offset; and

the monitoring sub-module processes the monitoring data acquired at the in-therapy monitoring stage of the clinical therapy process according to the monitoring model and outputs preliminary adjustment data.

2. The system of claim 1 , wherein:

the calibration database is further configured to receive modified sample data, and the modified sample data comprises modified data for the preliminary therapy data, and detection data corresponding to the preliminary therapy data, wherein the modified data is the data actually used in clinical therapy and obtained after clinicians modifying the preliminary therapy data; and

the training module is further configured to set a weight value of the modified sample data to be greater than a preset weight value, and perform deep learning on updated sample data in the calibration database.

3. The system of claim 1 , further comprising: a communication module, wherein:

the communication module is configured to send the therapy algorithm model to a cloud server, so that the cloud server sends the therapy algorithm model to local servers of other hospitals;

or,

the radiotherapy system is deployed in the cloud server, the plurality of sample data stored in the calibration database are acquired from a clinical database of a specified hospital, and the communication module is configured to send the therapy algorithm model to a local server of any hospital.

4. The system of claim 1 , wherein the radiotherapy system is deployed in a cloud server; a plurality of sample data stored in the calibration database is acquired from a clinical database of a specified hospital; and

the data processing module is configured to receive detection data sent from a local server of any hospital, process the detection data according to the therapy algorithm model to generate preliminary therapy data, and send the preliminary therapy data to the local server.

5. The system of claim 1 , wherein:

each of the sample data further comprises a set of additional data comprising at least one patient attribute information;

the data received by the data processing module in the clinical therapy process further comprises additional data corresponding to the detection data; and the data processing module is further configured to process the received detection data and the additional data corresponding to the detection data according to the therapy algorithm model to generate the preliminary therapy data.

6. The system of claim 1 , wherein:

each of the sample data further comprises a set of additional data comprising at least one patient attribute information;

the calibration database is further configured to classify the plurality of sample data according to content of target patient attribute information in the additional data of each sample data, the target patient attribute information is determined from the at least one patient attribute information;

the training module is configured to perform deep learning on each classification of sample data stored in the calibration database to obtain multiple classifications of therapy algorithm models;

the data received by the data processing module in the clinical therapy process further comprises additional data corresponding to the detection data; and the data processing module is further configured to determine a corresponding therapy algorithm model from the multiple classifications of therapy algorithm models according to the content of the target patient attribute information in the additional data, and process the received detection data according to the corresponding therapy algorithm model to generate the preliminary therapy data.

7. The system of claim 1 , wherein the clinical therapy process is used to treat the patient.

8. A data processing method applied to a radiotherapy system, comprising:

acquiring, by a calibration database which is accessed to a clinical database through a preset database interface, in real time or periodically a plurality of the sample data generated in a clinical therapy process;

performing, by a training module which is connected to the calibration database through a preset interface, deep learning on any one of the plurality of sample data stored in a calibration database to obtain a therapy algorithm model, and sending the therapy algorithm model to a data processing module;

processing, by the data processing module, the detection data received in the clinical therapy process according to the therapy algorithm model to generate preliminary therapy data;

wherein each sample data comprises a set of detection data and a set of therapy data corresponding to the set of the detection data and applied to the clinical therapy process and the therapy data is verified in the clinical therapy process;

the clinical therapy process comprises a therapy plan generation stage, a pre-therapy positioning stage and an in-therapy monitoring stage;

in the therapy plan generation stage, the detection data comprises medical image data, the therapy data comprises a therapy plan, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the therapy plan generation stage to obtain a therapy plan formulation model;

in the pre-therapy positioning stage, the detection data comprises medical image data acquired before the clinical therapy process and in the clinical therapy process, the therapy data comprises a positioning offset of a patient, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the pre-therapy positioning stage to obtain a pre-therapy positioning model;

in the in-therapy monitoring stage, the detection data comprises a patient displacement, a tumor monitoring image, a remaining radiation dosage, the therapy data comprises a tumor displacement before and after therapy, a dosage error, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the in-therapy monitoring stage to obtain a monitoring model;

the calibration database classifies and stores the detection data and the therapy data generated in the therapy plan generation stage, the pre-therapy positioning stage and the in- therapy monitoring stage;

receiving a feedback level of a specified sample data, wherein the feedback level is positively related to a therapeutic effect of therapy data in the specified sample data;

adjusting a weight value of the specified sample data according to the feedback level of the specified sample data, the adjusted weight value of the specified sample data being positively related to the feedback level of the specified sample data;

performing deep learning on sample data with the adjusted weight value; and

wherein:

the data processing module comprises an image registration sub-module, a delineation sub-module, a therapy plan formulation sub-module, a positioning sub-module and a monitoring sub-module;

the training module performs deep learning on a plurality of medical image data and registration images generated at an image registration stage to obtain a registration model; and the image registration sub-module processes medical image data acquired in the clinical therapy process according to the registration model and outputs a preliminary registration image;

the training module performs deep learning on the plurality of registration images and delineation result images generated at a delineation stage to obtain a delineation model; and the delineation sub-module processes registration images acquired in the clinical therapy process according to the delineation model and outputs a preliminary delineation result image;

the positioning sub-module processes the medical image data acquired at the image registration stage and the medical image data acquired at the pre-therapy positioning stage during the clinical therapy process according to the pre-therapy positioning model and outputs a preliminary positioning offset and

the monitoring sub-module processes the monitoring data acquired at the in-therapy monitoring stage of the clinical therapy process according to the monitoring model and outputs preliminary adjustment data.

9. The method of claim 8 , further comprising:

receiving modified sample data, wherein the modified sample data comprises modified data for the preliminary therapy data, and detection data corresponding to the preliminary therapy data, wherein the modified data is actually used in clinical therapy and obtained after clinicians modifying the preliminary therapy data;

setting a weight value of the modified sample data to be greater than a preset weight value; and

performing deep learning on updated sample data in the calibration database.

10. The method of claim 8 , wherein:

after obtaining the therapy algorithm model, the method further comprises: sending the therapy algorithm model to a cloud server, so that the cloud server sends the therapy algorithm model to local servers of other hospitals;

or,

the radiotherapy system is deployed in a cloud server; the acquiring the plurality of the sample data generated in the clinical therapy process comprises: acquiring the plurality of sample data from a clinical database of a specified hospital; and after obtaining the therapy algorithm model, the method further comprises: sending the therapy algorithm model to a local server of any hospital.

11. The method of claim 8 , wherein the radiotherapy system is deployed in a cloud server; the acquiring the plurality of the sample data generated in the clinical therapy process comprises:

acquiring the plurality of sample data from a clinical database of a specified hospital;

the processing the detection data received in the clinical therapy process according to the therapy algorithm model to generate the preliminary therapy data comprises:

receiving detection data sent from a local server of any hospital, and

processing the detection data according to the therapy algorithm model to generate the preliminary therapy data;

wherein the method further comprises: sending the preliminary therapy data to the local server.

12. The method of claim 8 , wherein:

each of the sample data further comprises a set of additional data comprising at least one patient attribute information;

the data received in the clinical therapy process further comprises additional data corresponding to the detection data; and the processing the detection data received in the clinical therapy process according to the therapy algorithm model to generate the preliminary therapy data comprises:

processing the received detection data and the corresponding additional data according to the therapy algorithm model to generate the preliminary therapy data.

13. The method of claim 8 , wherein:

each of the sample data further comprises a set of additional data comprising at least one patient attribute information; and the method further comprises:

classifying the plurality of sample data according to the content of target patient attribute information in the additional data of each sample data, the target patient attribute information being determined from the at least one patient attribute information;

the performing deep learning on the plurality of sample data stored in the calibration database to obtain a therapy algorithm model comprises:

performing deep learning on each classification of sample data stored in the calibration database to obtain multiple classifications of therapy algorithm models;

the data received in the clinical therapy process further comprises additional data corresponding to the detection data; and the processing the detection data received in the clinical therapy process according to the therapy algorithm model to generate the preliminary therapy data comprises:

determining a corresponding therapy algorithm model from the multiple classifications of therapy algorithm models according to the content of the target patient attribute information in the additional data; and

processing the received detection data according to the corresponding therapy algorithm model to generate the preliminary therapy data.

14. The method of claim 8 , further comprising treating the patient using the clinical therapy process.

15. A non-transitory computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium; and when the computer-readable storage medium runs on a computer, the computer executes a data processing method applied to a radiotherapy system, comprising:

acquiring, by a calibration database which is accessed to a clinical database through a preset database interface, in real time or periodically a plurality of the sample data generated in a clinical therapy process;

performing, by a training module which is connected to the calibration database through a preset interface, deep learning on any one of the plurality of sample data stored in a calibration database to obtain a therapy algorithm model, and sending the therapy algorithm model to a data processing module;

processing, by the data processing module, the detection data received in the clinical therapy process according to the therapy algorithm model to generate preliminary therapy data;

wherein each sample data comprises a set of detection data and a set of therapy data corresponding to the set of the detection data and applied to the clinical therapy process and the therapy data is verified in the clinical therapy;

the clinical therapy process comprises a therapy plan generation stage, a pre-therapy positioning stage and an in-therapy monitoring stage;

in the therapy plan generation stage, the detection data comprises medical image data, the therapy data comprises a therapy plan, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the therapy plan generation stage to obtain a therapy plan formulation model;

in the pre-therapy positioning stage, the detection data comprises medical image data acquired before the clinical therapy process and in the clinical therapy process, the therapy data comprises a positioning offset of a patient, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the pre-therapy positioning stage to obtain a pre-therapy positioning model;

in the in-therapy monitoring stage, the detection data comprises a patient displacement, a tumor monitoring image, a remaining radiation dosage, the therapy data comprises a tumor displacement before and after therapy, a dosage error, and the training module is configured to perform deep learning on the detection data and the therapy data generated in the in-therapy monitoring stage to obtain a monitoring model;

the calibration database classifies and stores the detection data and the therapy data generated in the therapy plan generation stage, the pre-therapy positioning stage and the in- therapy monitoring stage;

receiving a feedback level of a specified sample data, wherein the feedback level is positively related to a therapeutic effect of therapy data in the specified sample data;

adjusting a weight value of the specified sample data according to the feedback level of the specified sample data, the adjusted weight value of the specified sample data being positively related to the feedback level of the specified sample data;

performing deep learning on sample data with the adjusted weight value; and

wherein:

the data processing module comprises an image registration sub-module, a delineation sub-module, a therapy plan formulation sub-module, a positioning sub-module and a monitoring sub-module;

the training module performs deep learning on a plurality of medical image data and registration images generated at an image registration stage to obtain a registration model; and the image registration sub-module processes medical image data acquired in the clinical therapy process according to the registration model and outputs a preliminary registration image;

the training module performs deep learning on the plurality of registration images and delineation result images generated at a delineation stage to obtain a delineation model; and the delineation sub-module processes registration images acquired in the clinical therapy process according to the delineation model and outputs a preliminary delineation result image;

the positioning sub-module processes the medical image data acquired at the image registration stage and the medical image data acquired at the pre-therapy positioning stage during the clinical therapy process according to the pre-therapy positioning model and outputs a preliminary positioning offset and

the monitoring sub-module processes the monitoring data acquired at the in-therapy monitoring stage of the clinical therapy process according to the monitoring model and outputs preliminary adjustment data.

16. The computer-readable storage medium of claim 15 , further comprising:

receiving modified sample data, wherein the modified sample data comprises modified data for the preliminary therapy data, and detection data corresponding to the preliminary therapy data, wherein the modified data is actually used in clinical therapy and obtained after clinicians modifying the preliminary therapy data;

setting a weight value of the modified sample data to be greater than a preset weight value; and

performing deep learning on updated sample data in the calibration database.