CN117711552A - Radiotherapy clinical auxiliary decision-making method and system based on artificial intelligence knowledge base - Google Patents
Radiotherapy clinical auxiliary decision-making method and system based on artificial intelligence knowledge base Download PDFInfo
- Publication number
- CN117711552A CN117711552A CN202410165064.0A CN202410165064A CN117711552A CN 117711552 A CN117711552 A CN 117711552A CN 202410165064 A CN202410165064 A CN 202410165064A CN 117711552 A CN117711552 A CN 117711552A
- Authority
- CN
- China
- Prior art keywords
- data
- information
- radiotherapy
- key information
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001959 radiotherapy Methods 0.000 title claims abstract description 89
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 106
- 206010028980 Neoplasm Diseases 0.000 claims abstract description 63
- 238000012360 testing method Methods 0.000 claims abstract description 45
- 238000011156 evaluation Methods 0.000 claims abstract description 43
- 230000008901 benefit Effects 0.000 claims abstract description 42
- 238000012545 processing Methods 0.000 claims abstract description 37
- 238000003745 diagnosis Methods 0.000 claims abstract description 34
- 238000005516 engineering process Methods 0.000 claims abstract description 20
- 238000004140 cleaning Methods 0.000 claims description 39
- 238000011282 treatment Methods 0.000 claims description 37
- 238000013527 convolutional neural network Methods 0.000 claims description 21
- 230000005855 radiation Effects 0.000 claims description 19
- 238000007781 pre-processing Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 11
- 238000010801 machine learning Methods 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 3
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 230000007170 pathology Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000002068 genetic effect Effects 0.000 description 3
- 210000000920 organ at risk Anatomy 0.000 description 3
- 238000006467 substitution reaction Methods 0.000 description 3
- 230000004075 alteration Effects 0.000 description 2
- 230000008595 infiltration Effects 0.000 description 2
- 238000001764 infiltration Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 108090000623 proteins and genes Proteins 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011243 body radiation therapy Methods 0.000 description 1
- 238000002725 brachytherapy Methods 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013211 curve analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000004980 dosimetry Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000002710 external beam radiation therapy Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 238000011269 treatment regimen Methods 0.000 description 1
- 210000004881 tumor cell Anatomy 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Abstract
The invention discloses a radiotherapy clinical auxiliary decision-making method and a radiotherapy clinical auxiliary decision-making system based on an artificial intelligence knowledge base, wherein the method comprises the following steps: acquiring the existing histology information of the patient; according to the NLP technology, extracting key information related to tumor characteristics, performing multi-mode data processing on the extracted key information, splitting the key information into a training set and a testing set, training and optimizing a supervised learning algorithm training model, and outputting a target supervised learning algorithm training model; inputting the existing key information of the target patient into a target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and the specific information related to tumor diagnosis. The invention integrates an artificial intelligence NLP technology and a multi-mode processing technology, and obtains a recommended scheme of a radiotherapy scheme through the histology information of a patient, thereby making an auxiliary decision on a radiotherapy plan.
Description
Technical Field
The invention relates to the technical field of clinical decision knowledge base based on knowledge base, in particular to a radiotherapy clinical auxiliary decision making method and system based on an artificial intelligence knowledge base.
Background
Tumor radiotherapy is a treatment method which utilizes particle radiation to bombard tumors and simultaneously utilizes biological differences of tumor tissues and healthy tissues aiming at the radiation to kill the tumors and protect the healthy tissues.
The current tumor radiotherapy process comprises the following steps:
diagnostic examination: the first step in tumor radiotherapy, mainly to diagnose a tumor, determines the location, size and extent of the tumor, may require some imaging examinations such as CT, MRI, etc.
Making a treatment scheme: depending on the patient's condition, the radiologist determines an appropriate radiation treatment plan including external-beam radiation therapy, brachytherapy, or whole-body radiation therapy. The program will take into account the overall health of the patient, the type and stage of the cancer and the location of the tumor. The radiologist will also consider the potential side effects and complications of radiation therapy, as well as the treatment that the patient has previously received.
Positioning and target region delineation: prior to treatment, the physician needs to precisely locate the tumor and delineate the target area to be irradiated. This is to ensure that the radiation is accurately irradiated to the tumor site, thereby killing the tumor cells to the maximum.
Planning design, assessment and validation: the doctor can design the optimal radiotherapy scheme according to the specific condition of the patient. This scheme may involve the dose of irradiation, the number of times of irradiation, the manner of irradiation, etc. The physician then evaluates the regimen to see if it is able to achieve the desired therapeutic effect, and verifies the feasibility of the regimen.
Treatment is started: after verification of the treatment plan, radiation therapy can be started. This procedure may require multiple exposures, each of which may be dosed and counted according to the treatment regimen.
Discharge and follow-up: after the treatment is finished, the patient needs to be discharged from the hospital according to the advice of the doctor and follow-up visit is carried out. The doctor can check the physical condition of the patient regularly to observe the treatment effect, and meanwhile, the possible complications can be found and treated in time.
The current tumor radiotherapy process has the following points: traditional tumor radiation therapy procedures often rely on the physician's processing of image data and understanding memory of various guidelines, lacking in systemicity, timeliness and integrity.
Therefore, the invention provides a radiotherapy clinical auxiliary decision-making method and a radiotherapy clinical auxiliary decision-making system based on an artificial intelligence knowledge base.
Disclosure of Invention
Based on the above, it is necessary to provide a radiotherapy clinical auxiliary decision-making method and system based on an artificial intelligence knowledge base.
In order to achieve the above object, the technical scheme of the present invention is as follows:
in a first aspect, the present invention provides a radiotherapy clinical aid decision-making method based on an artificial intelligence knowledge base, the method comprising:
acquiring the existing histology information of the patient;
cleaning and preprocessing the histology information to obtain a histology information data set;
according to the NLP technology, carrying out information extraction processing on the histology information contained in the histology information data set, extracting key information related to tumor characteristics, carrying out multi-mode data processing on the extracted key information, integrating and analyzing data modes from different sources, splitting the key information processed by the multi-mode data into a training set and a testing set, training and optimizing a supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model;
inputting the existing key information of the target patient into a target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and the specific information related to tumor diagnosis.
In some embodiments, the performing the cleaning pretreatment on the omics information to obtain a omics information data set includes: presetting a data cleaning rule according to data cleaning requirements, performing traversal search on the group study information based on the preset data cleaning rule, and cleaning away wrong and repeated group study information when the traversal search does not accord with the data cleaning rule, so as to form a group study information data set.
In some embodiments, according to the NLP technique, the processing of extracting the histology information included in the histology information data set, extracting the key information related to the tumor characteristics, processing the extracted key information in a multi-mode data, integrating and analyzing data modes from different sources, splitting the key information processed in the multi-mode data into a training set and a testing set, training and optimizing the supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model, including:
according to the NLP technology, identifying and marking feature points related to tumors by screening relevant fields of the histology information contained in the histology information data set, and then dividing text data contained in the feature points into application fields of different sources to extract key information related to tumor characteristics;
performing multi-mode data processing on the extracted key information according to data modes from different sources, wherein the data modes comprise image data, unstructured data and structured data, when the data modes are the image data, the extracted key information is formed into a structured diagnosis report, when the data modes are the unstructured data, the extracted key information is converted into the structured data, and when the data modes are the structured data, the extracted key information is sequenced and integrated;
splitting the key information processed by the multi-mode data into a training set and a testing set, adopting the training set to carry out deep learning on the deep convolutional neural network model, adopting the testing set to test the deep convolutional neural network model, optimizing model parameters according to test results, and outputting a target supervised learning algorithm training model.
In some embodiments, the inputting the existing key information of the target patient into the target supervised learning algorithm training model to obtain the benefit evaluation result of the target patient for performing different radiotherapy schemes, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation result of the target patient and the specific information related to tumor diagnosis includes: inputting the existing key information of the target patient into a deep convolutional neural network model, outputting the benefit evaluation results of different radiation treatment schemes of the target patient, listing different radiation treatment schemes according to the preset score recommendation rules and descending score order according to the benefit evaluation results of the target patient and specific information related to tumor diagnosis, and preferentially obtaining recommended schemes from the specified number of radiation treatment schemes.
In some embodiments, further comprising: and carrying out machine learning on the target supervised learning algorithm training model in real time according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad so as to update and optimize the target supervised learning algorithm training model.
In a second aspect, the present invention provides a radiotherapy clinical aid decision making system based on an artificial intelligence knowledge base, the system comprising:
the data acquisition module is used for acquiring the existing histology information of the patient;
the data preprocessing module is used for cleaning and preprocessing the histology information to obtain a histology information data set;
the radiotherapy decision support system is used for carrying out information extraction processing on the histology information contained in the histology information data set according to the NLP technology, extracting key information related to tumor characteristics, carrying out multi-mode data processing on the extracted key information, integrating and analyzing data modes from different sources, splitting the key information processed by the multi-mode data into a training set and a testing set, training and optimizing a supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model;
the user interface module is used for inputting the existing key information of the target patient into the target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and the specific information related to tumor diagnosis.
In some embodiments, the data preprocessing module comprises:
the data cleaning rule definition module is used for presetting data cleaning rules according to data cleaning requirements;
the data detection analysis module is used for performing traversal search on the group information based on a preset data cleaning rule, and when the traversal search does not accord with the data cleaning rule, the wrong and repeated group information is cleaned out to form a group information data set.
In some embodiments, the radiotherapy decision support system comprises:
the key information extraction module is used for identifying and marking feature points related to the tumor by screening relevant fields of the histology information contained in the histology information data set according to an NLP technology, and then dividing text data contained in the feature points into application fields of different sources to extract key information related to the tumor characteristics;
the multi-mode data processing module is used for carrying out multi-mode data processing on the extracted key information according to data modes from different sources, wherein the data modes comprise image data, unstructured data and structured data, the extracted key information is formed into a structured diagnosis report when the data modes are the image data, the extracted key information is converted into the structured data when the data modes are the unstructured data, and the extracted key information is sequenced and integrated when the data modes are the structured data;
the supervised learning algorithm training model learning and optimizing module is used for splitting the key information processed by the multi-mode data into a training set and a testing set, performing deep learning on the deep convolutional neural network model by adopting the training set, testing the deep convolutional neural network model by adopting the testing set, optimizing model parameters according to the testing result, and outputting a target supervised learning algorithm training model.
In some embodiments, the user interface module comprises:
the patient benefit evaluation module is used for inputting the existing key information of the target patient into the deep convolutional neural network model and outputting benefit evaluation results of different radiotherapy schemes of the target patient;
and the radiotherapy scheme recommending module lists different radiotherapy schemes according to a grading descending order according to a preset grading recommending rule according to a benefited evaluation result of a target patient and specific information related to tumor diagnosis, and preferentially obtains recommended schemes from a specified number of radiotherapy schemes.
In some embodiments, further comprising: and the supervised learning algorithm training model updating and learning module is used for carrying out machine learning on the target supervised learning algorithm training model in real time according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad so as to update and optimize the target supervised learning algorithm training model.
The invention has the advantages that:
the radiotherapy clinical auxiliary decision-making method and system based on the artificial intelligence knowledge base provided by the invention have the following beneficial effects:
1. adaptivity: the recommended scheme of the radiotherapy scheme can be obtained according to the histology information of the patient, so that an auxiliary decision is made on the radiotherapy scheme;
2. safety: advice is provided through artificial intelligence auxiliary decision making, and a doctor rechecks to ensure the treatment safety;
3. high efficiency: the doctor does not need to spend a lot of time to collect related information, and the radiotherapy clinical auxiliary decision-making system automatically gives suggestions and reminders, so that the workload of medical staff is reduced;
4. high security: the radiation treatment scheme, the review node and the like are more personalized, and more volumes of patients are reduced to receive unnecessary dose irradiation;
5. accuracy: the method can update and learn according to the new data and feedback information, and effectively improve the accuracy of clinical auxiliary decision making of radiotherapy.
Drawings
FIG. 1 is a flow chart of a method of clinical aid decision making for radiation therapy based on an artificial intelligence knowledge base;
fig. 2 is a schematic structural diagram of an artificial intelligence knowledge base-based radiotherapy clinical aid decision making system.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail by the following detailed description with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Example 1
Referring to fig. 1, fig. 1 is a flow chart of an artificial intelligence knowledge base-based radiotherapy clinical auxiliary decision making method provided in this embodiment, and the method mainly includes the following steps:
s1, acquiring the existing histology information of the patient.
In this embodiment, the patient's existing histology information is collected from sources such as Hospital Information System (HIS), image storage communication system (PACS), pathology information system (LIS), etc., wherein the histology information includes, but is not limited to, clinical data, image data, pathology data, gene data, etc.
S2, cleaning and preprocessing the histology information to obtain a histology information data set.
In this embodiment, the cleaning pretreatment is performed on the omic information to obtain a omic information data set, including: presetting a data cleaning rule according to data cleaning requirements, performing traversal search on the group study information based on the preset data cleaning rule, and cleaning away wrong and repeated group study information to form a group study information data set when the traversal search does not accord with the data cleaning rule, wherein the cleaning preprocessing step is adopted for: error and duplicate information is eliminated in preparation for subsequent processing.
S3, carrying out information extraction processing on the histology information contained in the histology information data set according to an NLP technology, extracting key information related to tumor characteristics, carrying out multi-mode data processing on the extracted key information, integrating and analyzing data modes from different sources, splitting the key information subjected to multi-mode data processing into a training set and a testing set, training and optimizing a supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model.
In this embodiment, according to the NLP technique, the information extraction processing is performed on the omic information included in the omic information dataset, the key information related to the tumor characteristics is extracted, the multi-modal data processing is performed on the extracted key information, the data modalities from different sources are integrated and analyzed, the key information after the multi-modal data processing is split into a training set and a test set, the supervised learning algorithm training model is trained and optimized according to the training set and the test set, and the target supervised learning algorithm training model is output, including:
according to the NLP technology, identifying and marking feature points related to tumors by screening relevant fields of the histology information contained in the histology information data set, and then dividing text data contained in the feature points into application fields of different sources to extract key information related to tumor characteristics;
performing multi-mode data processing on the extracted key information according to data modes from different sources, wherein the data modes comprise image data, unstructured data and structured data, when the data modes are the image data, the extracted key information is formed into a structured diagnosis report, when the data modes are the unstructured data, the extracted key information is converted into the structured data, and when the data modes are the structured data, the extracted key information is sequenced and integrated;
splitting the key information processed by the multi-mode data into a training set and a testing set, adopting the training set to carry out deep learning on the deep convolutional neural network model, adopting the testing set to test the deep convolutional neural network model, optimizing model parameters according to test results, and outputting a target supervised learning algorithm training model.
In this embodiment, the radiotherapy decision support system (CDSS) of the present invention is a core part of the auxiliary decision system, and the part integrates a machine learning algorithm and an artificial intelligence technology, and performs deep learning and pattern recognition according to the patient's group learning information, so as to obtain an accurate training model of the target supervised learning algorithm.
In this embodiment, a large amount of text information is parsed and processed by using NLP technology, and key information related to tumor characteristics is extracted, wherein the text information includes, but is not limited to, medical records, images, inspection results, genetic data, reports, and the like of patients required in diagnosis and treatment.
In the present embodiment, for example: the NLP technology is used for analyzing and processing a large amount of text information, and key information related to tumor such as tumor size, position, infiltration degree and the like is extracted by analyzing medical records, images and other data (such as inspection and check, genetic data and the like) of patients.
In the embodiment, the supervised learning algorithm training model adopts the deep convolutional neural network model to analyze and decide the data, and simultaneously, a large amount of data can be stored and processed by using the deep convolutional neural network model, so that the computing and data processing capacity is effectively improved.
S4, inputting the existing key information of the target patient into a target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and specific information related to tumor diagnosis.
In this embodiment, the present key information of the target patient is input to a training model of a target supervised learning algorithm to obtain the benefit evaluation result of the target patient for different radiation treatment schemes, and the recommended scheme of the radiation treatment scheme is obtained according to the benefit evaluation result of the target patient and specific information related to tumor diagnosis, including: inputting the existing key information of the target patient into a deep convolutional neural network model, outputting the benefit evaluation results of different radiation treatment schemes of the target patient, listing different radiation treatment schemes according to the preset score recommendation rules and descending score order according to the benefit evaluation results of the target patient and specific information related to tumor diagnosis, and preferentially obtaining recommended schemes from the specified number of radiation treatment schemes.
In this embodiment, key information such as pathology and genetic information of a patient related to tumor characteristics extracted from a target patient in advance is input into a deep convolutional neural network model, so as to obtain benefit evaluation results of different radiotherapy schemes of the target patient, where the benefit evaluation results are comparison analysis results of different radiotherapy schemes, for example: by performing dosimetry comparison on one or more suggested radiotherapy schemes and manual design schemes, the parameters such as absolute dose, tame degree, organ-jeopardy dose constraint and the like of a target area are checked for evaluation, and corresponding benefit evaluation results are obtained.
In this embodiment, the system may provide the doctor with possible radiotherapy schemes and suggestions, including recommended schemes such as radiation source selection, irradiation field design, and dose division pattern, according to the result of the benefit evaluation of the target patient and specific information related to tumor diagnosis.
Specific information related to tumor diagnosis includes, but is not limited to, tumor stage, type, position size, infiltration degree and the like.
The system recommends rules according to preset scores according to the benefit evaluation result of the target patient and specific information related to tumor diagnosis, such as: the isodose curve analysis score and/or the DVH map analysis score and/or the dose uniformity analysis score and/or the dose limiting of the organ at risk are ranked from high to low, and a recommended plan is preferentially derived from a specified number of radiation treatment plans.
Furthermore, the isodose curve is a group of curves formed by connecting points with the same dose on a central axis plane of the overray in the die body, and the isodose curve can intuitively reflect the dose change of the ray beam in the in-vivo off-axis direction, and when the three-dimensional dose curve is analyzed, each index needs to meet each specification of the dose of the external irradiation target area, such as: the compliance index (the minimum dose of the target area, typically represented by 90% isodose lines, is taken as the lower limit of the range of the treatment area, the better the shape and size of the treatment area are in line with the planned target area, the better the plan), the size of the irradiated area (the irradiated area is the range included by 50% isodose line surfaces, the size of the irradiated area directly reflects the size of normal tissue dose), the dose hot spot and the cold spot area (the dose hot spot is defined as the range of the dose area which is larger than the specified target dose outside the inner target area, and the hot spot area is larger than or equal to 2cm2 (1.5 cm in diameter), and the influence is not considered clinically when the area is smaller than 2cm 2).
Further, the DVH image can show how much of the volume within the target and vital organ is irradiated by the high dose level, and the suitability of the high dose area to the target can be directly assessed according to the DVH image, and a better treatment plan is to gauge the dose (100%) at the target site by 100% of the volume of the receptor, while compromising the 100% of the volume of the receptor to 0, but in practice it is difficult to achieve, so that 95% of the volume is required; radiation treatment plans in the PTV that receive a prescribed dose volume greater than 95% of the PTV volume are clinically acceptable.
Further, the dose uniformity rule is that the deviation from the prescribed dose for the dose distribution of the intended target is preferably no more than-5% to +7%.
Further, the dose rule for the organ at risk is that the organ at risk within and around the target area is to meet a prescribed dose limit.
And S5, performing machine learning on the target supervised learning algorithm training model in real time according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad so as to update and optimize the target supervised learning algorithm training model.
In the embodiment, the machine learning is performed on the target supervised learning algorithm training model by acquiring the latest clinical evidences such as diagnosis and treatment guidelines, consensus and the like issued by authoritative academic institutions, academic journals and conferences at home and abroad so as to update and optimize the target supervised learning algorithm training model, so that the accuracy of decision making can be improved.
In addition, the conditions of disease treatment effect, middle-long-term complications and the like can be detected through long-term follow-up of patients in clinical experiments, and the information is supplemented into medical records of the patients, so that the training model of the target supervision learning algorithm is continuously trained and optimized.
The radiotherapy clinical auxiliary decision-making method based on the artificial intelligence knowledge base provided by the embodiment comprises the following steps: acquiring the existing histology information of the patient; cleaning and preprocessing the histology information to obtain a histology information data set; according to the NLP technology, carrying out information extraction processing on the histology information contained in the histology information data set, extracting key information related to tumor characteristics, carrying out multi-mode data processing on the extracted key information, integrating and analyzing data modes from different sources, splitting the key information processed by the multi-mode data into a training set and a testing set, training and optimizing a supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model; inputting the existing key information of the target patient into a target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and the specific information related to tumor diagnosis; according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad, machine learning is carried out on the target supervised learning algorithm training model in real time so as to update and optimize the target supervised learning algorithm training model; the implementation method is characterized in that clinical, image, pathology, gene and other genomic information related to radiotherapy are integrated, biological characteristics of tumors are mined through an artificial intelligence NLP technology and a multi-mode processing technology, whether a patient can benefit from radiotherapy or not is judged, a recommended scheme of a radiotherapy scheme is obtained, an auxiliary decision is made on a radiotherapy plan, self-updating and learning can be carried out according to new data and feedback information, and accuracy of the radiotherapy clinical auxiliary decision is effectively improved.
Example two
On the basis of the first embodiment, the present embodiment provides an artificial intelligence knowledge base-based radiotherapy clinical auxiliary decision-making system, please refer to fig. 2, for implementing the steps of the artificial intelligence knowledge base-based radiotherapy clinical auxiliary decision-making method described in the first embodiment, the system mainly includes:
a data acquisition module 10 for acquiring the existing histology information of the patient;
the data preprocessing module 20 is used for performing cleaning preprocessing on the omics information to obtain a omics information data set;
the radiotherapy decision support system 30 is configured to perform information extraction processing on the omic information included in the omic information dataset according to the NLP technology, extract key information related to tumor characteristics, perform multi-modal data processing on the extracted key information, integrate and analyze data modalities from different sources, split the key information processed by the multi-modal data into a training set and a testing set, train and optimize a supervised learning algorithm training model according to the training set and the testing set, and output a target supervised learning algorithm training model;
the user interface module 40 is configured to input existing key information of the target patient into the target supervised learning algorithm training model to obtain a benefit evaluation result of the target patient for performing different radiotherapy schemes, and obtain a recommended scheme of the radiotherapy scheme according to the benefit evaluation result of the target patient and specific information related to tumor diagnosis.
In this embodiment, the data preprocessing module 20 includes:
the data cleaning rule definition module 201 is configured to preset a data cleaning rule according to a data cleaning requirement;
the data detection and analysis module 202 is configured to perform a traversal search on the group information based on a preset data cleansing rule, and cleanse the wrong and repeated group information to form a group information data set when the traversal search does not conform to the data cleansing rule.
In this embodiment, the radiotherapy decision support system 30 includes:
the key information extraction module 301 is configured to identify and mark feature points related to a tumor by screening relevant fields of the omic information included in the omic information dataset according to an NLP technique, and then divide text data included in the feature points into application fields of different sources to extract key information related to a tumor characteristic;
the multi-mode data processing module 302 is configured to perform multi-mode data processing on the extracted key information according to data modes from different sources, where the data modes include image data, unstructured data and structured data, form a structured diagnostic report on the extracted key information when the data mode is the image data, convert the extracted key information into the structured data when the data mode is the unstructured data, and sequence and integrate the extracted key information when the data mode is the structured data;
the supervised learning algorithm training model learning and optimizing module 303 is configured to split the key information processed by the multimodal data into a training set and a test set, perform deep learning on the deep convolutional neural network model by using the training set, test the deep convolutional neural network model by using the test set, optimize model parameters according to the test result, and output a target supervised learning algorithm training model.
In this embodiment, the user interface module 40 includes:
the patient benefit evaluation module 401 is configured to input existing key information of the target patient into the deep convolutional neural network model, and output benefit evaluation results of the target patient for different radiotherapy schemes;
the radiotherapy plan recommendation module 402 lists different radiotherapy plans according to a preset grading recommendation rule according to a grading descending order according to a benefited evaluation result of a target patient and specific information related to tumor diagnosis, and preferentially obtains recommended plans from a specified number of radiotherapy plans.
In this embodiment, further comprising: the supervised learning algorithm training model updating and learning module 50 is configured to machine learn the target supervised learning algorithm training model in real time according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad to update and optimize the target supervised learning algorithm training model.
Those skilled in the art will appreciate that while some embodiments described herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the present application and form different embodiments.
Those skilled in the art will appreciate that the descriptions of the various embodiments are each focused on, and that portions of one embodiment that are not described in detail may be referred to as related descriptions of other embodiments.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and alterations without departing from the spirit and scope of the present invention, and such modifications and alterations fall within the scope of the appended claims, which are to be construed as merely illustrative of this invention, but the scope of the invention is not limited thereto, and various equivalent modifications and substitutions will be readily apparent to those skilled in the art within the scope of the present invention, and are intended to be included within the scope of the invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
The present invention is not limited to the above embodiments, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the present invention, and these modifications and substitutions are intended to be included in the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (10)
1. The radiotherapy clinical auxiliary decision-making method based on the artificial intelligence knowledge base is characterized by comprising the following steps of:
acquiring the existing histology information of the patient;
cleaning and preprocessing the histology information to obtain a histology information data set;
according to the NLP technology, carrying out information extraction processing on the histology information contained in the histology information data set, extracting key information related to tumor characteristics, carrying out multi-mode data processing on the extracted key information, integrating and analyzing data modes from different sources, splitting the key information processed by the multi-mode data into a training set and a testing set, training and optimizing a supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model;
inputting the existing key information of the target patient into a target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and the specific information related to tumor diagnosis.
2. The method for clinical aid decision making of radiotherapy based on artificial intelligence knowledge base according to claim 1, wherein the step of performing a cleaning pretreatment on the omics information to obtain a omics information data set comprises the steps of: presetting a data cleaning rule according to data cleaning requirements, performing traversal search on the group study information based on the preset data cleaning rule, and cleaning away wrong and repeated group study information when the traversal search does not accord with the data cleaning rule, so as to form a group study information data set.
3. The method for clinical aid decision making of radiotherapy based on artificial intelligence knowledge base according to claim 1, wherein the method for clinical aid decision making comprises extracting information from the group information included in the group information dataset, extracting key information related to tumor characteristics, performing multi-modal data processing on the extracted key information, integrating and analyzing data modalities from different sources, splitting the key information processed by the multi-modal data into a training set and a test set, training and optimizing a supervised learning algorithm training model according to the training set and the test set, and outputting a target supervised learning algorithm training model, and comprises:
according to the NLP technology, identifying and marking feature points related to tumors by screening relevant fields of the histology information contained in the histology information data set, and then dividing text data contained in the feature points into application fields of different sources to extract key information related to tumor characteristics;
performing multi-mode data processing on the extracted key information according to data modes from different sources, wherein the data modes comprise image data, unstructured data and structured data, when the data modes are the image data, the extracted key information is formed into a structured diagnosis report, when the data modes are the unstructured data, the extracted key information is converted into the structured data, and when the data modes are the structured data, the extracted key information is sequenced and integrated;
splitting the key information processed by the multi-mode data into a training set and a testing set, adopting the training set to carry out deep learning on the deep convolutional neural network model, adopting the testing set to test the deep convolutional neural network model, optimizing model parameters according to test results, and outputting a target supervised learning algorithm training model.
4. The method for assisting in clinical decision making of radiotherapy based on artificial intelligence knowledge base according to claim 3, wherein the step of inputting the existing key information of the target patient into the training model of the target supervised learning algorithm to obtain the benefit evaluation result of the target patient for different radiotherapy schemes, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation result of the target patient and the specific information related to tumor diagnosis comprises the following steps: inputting the existing key information of the target patient into a deep convolutional neural network model, outputting the benefit evaluation results of different radiation treatment schemes of the target patient, listing different radiation treatment schemes according to the preset score recommendation rules and descending score order according to the benefit evaluation results of the target patient and specific information related to tumor diagnosis, and preferentially obtaining recommended schemes from the specified number of radiation treatment schemes.
5. The method for clinical aid decision making for radiation therapy based on artificial intelligence knowledge base according to claim 1, further comprising: and carrying out machine learning on the target supervised learning algorithm training model in real time according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad so as to update and optimize the target supervised learning algorithm training model.
6. A radiotherapy clinical aid decision making system based on an artificial intelligence knowledge base, comprising:
the data acquisition module is used for acquiring the existing histology information of the patient;
the data preprocessing module is used for cleaning and preprocessing the histology information to obtain a histology information data set;
the radiotherapy decision support system is used for carrying out information extraction processing on the histology information contained in the histology information data set according to the NLP technology, extracting key information related to tumor characteristics, carrying out multi-mode data processing on the extracted key information, integrating and analyzing data modes from different sources, splitting the key information processed by the multi-mode data into a training set and a testing set, training and optimizing a supervised learning algorithm training model according to the training set and the testing set, and outputting a target supervised learning algorithm training model;
the user interface module is used for inputting the existing key information of the target patient into the target supervised learning algorithm training model to obtain the benefit evaluation results of different radiotherapy schemes of the target patient, and obtaining the recommended scheme of the radiotherapy scheme according to the benefit evaluation results of the target patient and the specific information related to tumor diagnosis.
7. The radiation therapy clinical decision-making aid system based on artificial intelligence knowledge base according to claim 6, wherein the data preprocessing module comprises:
the data cleaning rule definition module is used for presetting data cleaning rules according to data cleaning requirements;
the data detection analysis module is used for performing traversal search on the group information based on a preset data cleaning rule, and when the traversal search does not accord with the data cleaning rule, the wrong and repeated group information is cleaned out to form a group information data set.
8. The artificial intelligence knowledge base based radiotherapy clinical aid decision making system of claim 6, wherein the radiotherapy decision support system comprises:
the key information extraction module is used for identifying and marking feature points related to the tumor by screening relevant fields of the histology information contained in the histology information data set according to an NLP technology, and then dividing text data contained in the feature points into application fields of different sources to extract key information related to the tumor characteristics;
the multi-mode data processing module is used for carrying out multi-mode data processing on the extracted key information according to data modes from different sources, wherein the data modes comprise image data, unstructured data and structured data, the extracted key information is formed into a structured diagnosis report when the data modes are the image data, the extracted key information is converted into the structured data when the data modes are the unstructured data, and the extracted key information is sequenced and integrated when the data modes are the structured data;
the supervised learning algorithm training model learning and optimizing module is used for splitting the key information processed by the multi-mode data into a training set and a testing set, performing deep learning on the deep convolutional neural network model by adopting the training set, testing the deep convolutional neural network model by adopting the testing set, optimizing model parameters according to the testing result, and outputting a target supervised learning algorithm training model.
9. The radiation therapy clinical aid decision making system based on artificial intelligence knowledge base according to claim 8, wherein the user interface module comprises:
the patient benefit evaluation module is used for inputting the existing key information of the target patient into the deep convolutional neural network model and outputting benefit evaluation results of different radiotherapy schemes of the target patient;
and the radiotherapy scheme recommending module lists different radiotherapy schemes according to a grading descending order according to a preset grading recommending rule according to a benefited evaluation result of a target patient and specific information related to tumor diagnosis, and preferentially obtains recommended schemes from a specified number of radiotherapy schemes.
10. The artificial intelligence knowledge base based radiotherapy clinical aid decision making system of claim 6, further comprising: and the supervised learning algorithm training model updating and learning module is used for carrying out machine learning on the target supervised learning algorithm training model in real time according to the latest authoritative clinical guidelines, diagnosis and treatment specifications, textbooks, documents, expert consensus and clinical evidence published by authoritative academic institutions, academic journals and conferences at home and abroad so as to update and optimize the target supervised learning algorithm training model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410165064.0A CN117711552A (en) | 2024-02-05 | 2024-02-05 | Radiotherapy clinical auxiliary decision-making method and system based on artificial intelligence knowledge base |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410165064.0A CN117711552A (en) | 2024-02-05 | 2024-02-05 | Radiotherapy clinical auxiliary decision-making method and system based on artificial intelligence knowledge base |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117711552A true CN117711552A (en) | 2024-03-15 |
Family
ID=90148310
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410165064.0A Pending CN117711552A (en) | 2024-02-05 | 2024-02-05 | Radiotherapy clinical auxiliary decision-making method and system based on artificial intelligence knowledge base |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117711552A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050049497A1 (en) * | 2003-06-25 | 2005-03-03 | Sriram Krishnan | Systems and methods for automated diagnosis and decision support for breast imaging |
| US20200069973A1 (en) * | 2018-05-30 | 2020-03-05 | Siemens Healthcare Gmbh | Decision Support System for Individualizing Radiotherapy Dose |
| WO2023232762A1 (en) * | 2022-05-30 | 2023-12-07 | Sophia Genetics Sa | Machine learning predictive models of treatment response |
-
2024
- 2024-02-05 CN CN202410165064.0A patent/CN117711552A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050049497A1 (en) * | 2003-06-25 | 2005-03-03 | Sriram Krishnan | Systems and methods for automated diagnosis and decision support for breast imaging |
| US20200069973A1 (en) * | 2018-05-30 | 2020-03-05 | Siemens Healthcare Gmbh | Decision Support System for Individualizing Radiotherapy Dose |
| WO2023232762A1 (en) * | 2022-05-30 | 2023-12-07 | Sophia Genetics Sa | Machine learning predictive models of treatment response |
Non-Patent Citations (4)
| Title |
|---|
| GILMER VALDES: "Clinical decision support of radiotherapy treatment planning: a data-driven machine learning strategy for patient-specific dosimetric decision making", RADIOTHERAPY AND ONCOLOGY, 31 December 2017 (2017-12-31), pages 392 - 397 * |
| NEHA KULKARNI; BHAVIN PATANWADIA; VIKRAM KULKARNI: "A Survey on Machine Learning Techniques for Breast Cancer Diagnosis and Detection", IEEE, 9 March 2022 (2022-03-09) * |
| 史佳: "基于深度学习的儿科临床辅助诊断算法研究", 优秀硕士学位论文, 15 September 2019 (2019-09-15), pages 3 - 35 * |
| 罗浩轩: "基于知识图谱的肿瘤决策支持系统设计与实现", 优秀硕士学位论文, 15 January 2023 (2023-01-15) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2719922C2 (en) | Adaptive treatment management system with workflow management engine | |
| US7606405B2 (en) | Dynamic tumor diagnostic and treatment system | |
| RU2684173C2 (en) | System and method for automatic treatment planning | |
| RU2529381C2 (en) | Formation of improved image model | |
| US20070156453A1 (en) | Integrated treatment planning system | |
| WO2018119766A1 (en) | Multi-modal image processing system and method | |
| EP3334496B1 (en) | Result-driven radiation therapy treatment planning | |
| US20120027271A1 (en) | Knowledge-based automatic image segmentation | |
| EP3013418B1 (en) | Portal dosimetry system | |
| WO2006119340A2 (en) | Dynamic tumor diagnostic and treatment system | |
| WO2019161135A1 (en) | Result-driven radiation therapy treatment planning | |
| US20120027273A1 (en) | Knowledge-based automatic image segmentation | |
| CN114707742A (en) | Artificial intelligence prediction method and system for adaptive radiotherapy strategy | |
| KR101968671B1 (en) | Intelligent automated method and system for setting radiation therapy plan | |
| CN110021399A (en) | A kind of automatic design method of radiotherapy treatment planning | |
| US10769240B2 (en) | Determining medical outcome quality | |
| CN114068001A (en) | Auxiliary support system and method for clinical decision, teaching and scientific research of head and neck tumors | |
| CN117711552A (en) | Radiotherapy clinical auxiliary decision-making method and system based on artificial intelligence knowledge base | |
| CN113196411A (en) | System for assisting in providing template treatment parameters for ablation treatment | |
| US20180169437A1 (en) | Result-driven radiation therapy treatment planning | |
| US8897417B2 (en) | Method and device for planning a treatment | |
| JP7330665B2 (en) | Treatment planning device and clinical model comparison method | |
| CN114067996A (en) | Nasopharyngeal carcinoma clinical decision, teaching and scientific research auxiliary support system and method | |
| CN101632604A (en) | Method and device for calculating a parameter affecting the prostate of a patient | |
| Liu | A knowledge-based imaging informatics approach for managing proton beam therapy of cancer patients |
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 |