CN111265234A - Method and system for judging properties of lung mediastinal lymph nodes - Google Patents
Method and system for judging properties of lung mediastinal lymph nodes Download PDFInfo
- Publication number
- CN111265234A CN111265234A CN202010148269.XA CN202010148269A CN111265234A CN 111265234 A CN111265234 A CN 111265234A CN 202010148269 A CN202010148269 A CN 202010148269A CN 111265234 A CN111265234 A CN 111265234A
- Authority
- CN
- China
- Prior art keywords
- image
- lung
- mediastinal lymph
- lymph node
- data
- 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
- 210000004072 lung Anatomy 0.000 title claims abstract description 72
- 210000005015 mediastinal lymph node Anatomy 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 22
- 201000011510 cancer Diseases 0.000 claims abstract description 53
- 206010028980 Neoplasm Diseases 0.000 claims abstract description 50
- 230000036210 malignancy Effects 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 47
- 230000003211 malignant effect Effects 0.000 claims abstract description 30
- 238000003325 tomography Methods 0.000 claims abstract description 29
- 238000003709 image segmentation Methods 0.000 claims abstract description 17
- 238000003759 clinical diagnosis Methods 0.000 claims abstract description 16
- 238000011156 evaluation Methods 0.000 claims description 37
- 230000002685 pulmonary effect Effects 0.000 claims description 20
- 230000003902 lesion Effects 0.000 claims description 18
- 238000012360 testing method Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 230000010354 integration Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 abstract description 4
- 230000000875 corresponding effect Effects 0.000 description 20
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 7
- 201000005202 lung cancer Diseases 0.000 description 7
- 208000020816 lung neoplasm Diseases 0.000 description 7
- 210000001165 lymph node Anatomy 0.000 description 7
- 238000002591 computed tomography Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000003745 diagnosis Methods 0.000 description 3
- 230000001575 pathological effect Effects 0.000 description 3
- 238000004393 prognosis Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000002512 chemotherapy Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000002059 diagnostic imaging Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 206010006187 Breast cancer Diseases 0.000 description 1
- 208000026310 Breast neoplasm Diseases 0.000 description 1
- 241001237259 Campanella <basidiomycete fungus> Species 0.000 description 1
- 208000007433 Lymphatic Metastasis Diseases 0.000 description 1
- 206010060862 Prostate cancer Diseases 0.000 description 1
- 208000000236 Prostatic Neoplasms Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 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
- 238000005516 engineering process Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 206010020718 hyperplasia Diseases 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002757 inflammatory effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 210000005259 peripheral blood Anatomy 0.000 description 1
- 239000011886 peripheral blood Substances 0.000 description 1
- 238000002600 positron emission tomography Methods 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000002271 resection Methods 0.000 description 1
- 210000005005 sentinel lymph node Anatomy 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 210000000115 thoracic cavity Anatomy 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/02—Arrangements for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
- A61B6/03—Computed tomography [CT]
- A61B6/032—Transmission computed tomography [CT]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/50—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Veterinary Medicine (AREA)
- General Physics & Mathematics (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- Biophysics (AREA)
- High Energy & Nuclear Physics (AREA)
- Surgery (AREA)
- Heart & Thoracic Surgery (AREA)
- Optics & Photonics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Pulmonology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Dentistry (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention provides a system and a method for judging the property of a lung mediastinal lymph node, which relate to the technical field of medical image processing and comprise the following steps: carrying out image segmentation on a plurality of acquired lung tomography images to obtain a focus position image of each lung tomography image; extracting high-dimensional characteristic data of each focus position image; acquiring the real malignant probability of the lung mediastinal lymph nodes aiming at each focus position image, and adding the focus position image, high-dimensional characteristic data and the real malignant probability into a focus data set; constructing an image characteristic database according to each focus data set; training according to the image feature database to obtain a property judgment model of the lung mediastinal lymph node; and inputting the lung tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignancy probability of the lung mediastinal lymph node of the patient to be predicted so as to be used by a doctor for clinical diagnosis reference. The accuracy of judging the properties of the mediastinal lymph nodes is effectively improved; the workload of doctors is reduced.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a method and a system for judging the property of a lung mediastinal lymph node.
Background
Lung cancer is one of the common malignant tumors in the world and is the leading cause of death related to cancer worldwide, the incidence and the death rate related to cancer are located at the first of the malignant tumors, about 220 ten thousand new cases are globally observed every year, and about 150 ten thousand death cases are observed; in China, the incidence rate and the cancer-related mortality of lung cancer still dominate malignant tumors, and the health of the nation is seriously threatened. And accurate judgment of the benign and malignant mediastinal lymph nodes of the lung cancer patient can help a clinician to judge and decide the cleaning range of the operation, assist chemotherapy and carry out survival prognosis evaluation, and has important significance for the stage, treatment and prognosis of the lung cancer.
At present, most of conventional examination, treatment scheme selection, treatment efficacy detection and the like of lung cancer patients are completed through medical images, and doctors clearly see internal focuses of the patients and monitor disease changes of the patients according to medical imaging examination methods such as Computed Tomography (CT) and Positron Emission Tomography (PET). In conventional diagnosis, a specialist often needs to determine the malignancy or malignancy of a lymph node by empirically comparing a series of images of an analyzed case. The method needs a professional doctor to perform complicated manual operation on a large amount of data, meanwhile, the accuracy and reliability of the diagnosis result of the method depend on the experience knowledge and professional quality of the doctor seriously, great uncertainty and difference exist, the evaluation value of the method on the lung cancer patient is very limited, and information such as pathological classification, benign and malignant lymph node and the like can not be obtained from the lung cancer focus image of the patient easily.
With the development of machine learning algorithms and computer hardware, deep learning (such as convolutional neural network algorithms, cyclic neural network algorithms, recurrent neural network algorithms, etc.) has begun to be applied to the processing of medical images. How to process a CT sequence image by a computer technology to complete tracking and identification of a lesion tissue instead of a doctor is a problem to be solved at present. In the field of thoracic medical imaging, in 2016, the CNN model is applied to detection of sentinel lymph nodes in a prostate cancer patient surgical resection specimen by Litjens and the like, so that the improvement of the diagnosis efficiency of the CNN model is discussed; in 2019, Campanella and the like also prove that the breast cancer lymph node metastasis cells can be accurately identified by using a composite neural network model consisting of CNN, MIL and RNN under the condition of weak supervised learning, and the accuracy can reach 0.965; however, there is no method for discriminating benign and malignant pulmonary mediastinal lymph nodes from CT images of the lung by using deep learning.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a property judgment system for a lung mediastinal lymph node, which specifically comprises the following steps:
the image segmentation module is used for acquiring lung tomography images of a plurality of patients and performing image segmentation on each lung tomography image to obtain a lesion site image of the lung mediastinal lymph node corresponding to each lung tomography image;
the characteristic extraction module is connected with the image segmentation module and used for respectively extracting the characteristics of the images of the focus parts to obtain high-dimensional characteristic data of the images of the focus parts;
the data integration module is respectively connected with the image segmentation module and the feature extraction module and is used for acquiring the real malignant probability of the lung mediastinal lymph nodes corresponding to the focus position image aiming at each focus position image and adding the focus position image, the high-dimensional feature data and the real malignant probability into a focus data set;
the database construction module is connected with the data integration module and used for constructing an image characteristic database according to each focus data set;
the model training module is connected with the database construction module and used for obtaining a property judgment model of the lung mediastinal lymph node according to the image feature database;
and the model prediction module is connected with the model training module and used for inputting the pulmonary tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignant probability of the pulmonary mediastinal lymph node of the patient to be predicted so as to be referred to clinical diagnosis by a doctor.
Preferably, the high-dimensional feature data includes morphological feature data, texture feature data, and/or grayscale feature data of the lesion site image.
Preferably, the true malignancy probability is 0 or 1.
Preferably, the model training module specifically includes:
the data dividing unit is used for dividing each focus data set in the image feature database into a training set and a testing set;
the data training unit is connected with the data dividing unit and used for training according to each focus data set in the training set to obtain a property judgment model of the lung mediastinal lymph node;
the model evaluation unit is respectively connected with the data dividing unit and the data training unit and used for inputting each focus data set in the test set into the property judgment model to obtain corresponding predicted malignancy probability, comparing each predicted malignancy probability with the corresponding real malignancy probability, and carrying out efficiency evaluation on the property judgment model according to the comparison result to obtain an efficiency evaluation result;
and the data output unit is respectively connected with the data training unit and the model evaluation unit and is used for outputting the property judgment model and the performance evaluation result, and the doctor evaluates the accuracy of the predicted malignancy probability of the property judgment model according to the performance evaluation result and uses the evaluation result as a clinical diagnosis reference.
A method for judging the property of a pulmonary mediastinal lymph node is applied to any one of the systems for judging the property of the pulmonary mediastinal lymph node, and the method for judging the property specifically comprises the following steps:
step S1, collecting lung tomography images of a plurality of patients, and carrying out image segmentation on the lung tomography images to obtain focus position images of lung mediastinal lymph nodes corresponding to the lung tomography images;
step S2, respectively carrying out feature extraction on each focus position image to obtain high-dimensional feature data of each focus position image;
step S3, for each focus position image, obtaining a true malignancy probability of the lung mediastinal lymph node corresponding to the focus position image, and adding the focus position image, the high-dimensional feature data and the true malignancy probability into a focus data set;
step S4, constructing an image characteristic database according to each focus data set;
step S5, training according to the image feature database to obtain a property judgment model of the lung mediastinal lymph node;
step S6, inputting the lung tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignancy probability of the lung mediastinal lymph node of the patient to be predicted, so that doctors can make reference for clinical diagnosis.
Preferably, in step S2, the high-dimensional feature data includes morphological feature data, texture feature data, and/or grayscale feature data of the lesion site image.
Preferably, in step S3, the true malignancy probability is 0 or 1.
Preferably, the step S5 specifically includes:
step S51, dividing each focus data set in the image feature database into a training set and a testing set;
step S52, training according to each focus data set in the training set to obtain a property judgment model of the lung mediastinal lymph node;
step S53, inputting each focus data set in the test set into the property judgment model to obtain corresponding predicted malignancy probability, comparing each predicted malignancy probability with the corresponding real malignancy probability, and performing efficacy evaluation on the property judgment model according to the comparison result to obtain an efficacy evaluation result;
and step S54, outputting the property judgment model and the performance evaluation result, and evaluating the accuracy of the predicted malignancy probability of the property judgment model by the doctor according to the performance evaluation result and using the result as a clinical diagnosis reference.
The technical scheme has the following advantages or beneficial effects:
1) the clinical real malignant probability of the patient is combined with the high-dimensional characteristics of the lung tomography image, so that the training of the property judgment model is performed, the accuracy of the property judgment of the mediastinal lymph node is effectively improved, and the method is noninvasive for the patient;
2) the predicted malignancy probability obtained by the property judgment model can provide effective clinical diagnosis reference for doctors, effectively reduces the workload of the doctors, and avoids uncertainty and difference caused by subjective judgment.
Drawings
FIG. 1 is a schematic diagram of a system for determining the characteristics of a mediastinal lymph node of the lung in accordance with a preferred embodiment of the present invention
FIG. 2 is a flow chart of a method for determining a property of a mediastinal lymph node of a lung according to a preferred embodiment of the present invention;
FIG. 3 is a flow chart illustrating a model training process according to a preferred embodiment of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present invention is not limited to the embodiment, and other embodiments may be included in the scope of the present invention as long as the gist of the present invention is satisfied.
In a preferred embodiment of the present invention, based on the above problems in the prior art, there is provided a system for determining a property of a pulmonary mediastinal lymph node, as shown in fig. 1, which specifically includes:
the image segmentation module 1 is used for acquiring lung tomography images of a plurality of patients and performing image segmentation on each lung tomography image to obtain a lesion site image of the lung mediastinal lymph node corresponding to each lung tomography image;
the feature extraction module 2 is connected with the image segmentation module 1 and is used for respectively extracting features of each focus position image to obtain high-dimensional feature data of each focus position image;
the data integration module 3 is respectively connected with the image segmentation module 1 and the feature extraction module 2 and is used for acquiring the real malignant probability of the lung mediastinal lymph node corresponding to the lesion site image aiming at each lesion site image and adding the lesion site image, the high-dimensional feature data and the real malignant probability into a lesion data set;
the database construction module 4 is connected with the data integration module 3 and is used for constructing an image characteristic database according to each focus data set;
the model training module 5 is connected with the database construction module 4 and used for obtaining a property judgment model of the lung mediastinal lymph node according to the training of the image feature database;
and the model prediction module 6 is connected with the model training module 5 and is used for inputting the lung tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignant probability of the lung mediastinal lymph node of the patient to be predicted so as to be referred to clinical diagnosis by a doctor.
Specifically, in this embodiment, the determination of benign or malignant lymph nodes at mediastinal lung can help clinicians to determine and decide the cleaning range of operation, whether to assist chemotherapy, and perform prognosis evaluation, the conventional image is generally determined by macroscopic features such as lymph node size, shape, boundary, and peripheral blood vessel condition, and the interpretation is highly subjective, there are observer differences, and there may be interference of lymph node inflammatory reactive hyperplasia, which all result in a great difference between the preoperative imaging data staging and the postoperative pathological staging of lymph nodes. Meanwhile, the clinical characteristics of the patient have more predictive significance in identifying the benign and malignant pulmonary lymph nodes, but the existing lesion analysis method only judges the benign and malignant pulmonary mediastinal lymph nodes through the image characteristics, because the image processing cannot acquire the clinical characteristics of the patient, particularly the characteristics of the patient who is treated or examined.
Therefore, the invention can accurately acquire the focus position image of the mediastinal lymph node by carrying out image segmentation on the lung tomography image, and avoids the interference caused by other background images. After obtaining a plurality of focus position images, respectively performing high-dimensional feature extraction on each focus position image to obtain high-dimensional feature data capable of representing the properties of mediastinal lymph nodes, wherein the high-dimensional feature data comprises but is not limited to morphological feature data, texture feature data and gray feature data of the mediastinal lymph nodes in the focus position images. And then, the real malignant probability of the lung mediastinal lymph node corresponding to the focus position image can be obtained through clinical pathological information of the patient recorded by the hospital, the value of the real malignant probability is 0 or 1, and preferably, when the lung mediastinal lymph node of the focus position image is benign, the value of the real malignant probability is 0, and when the lung mediastinal lymph node of the focus position image is malignant, the value of the real malignant probability is 1.
After the high-dimensional feature data and the true malignancy probability corresponding to each lesion site image are obtained, the lesion site images, the corresponding high-dimensional feature data and the true malignancy probability are preferably added into a lesion data set to construct an image feature database taking the lesion data set as a subset as a whole. And then training according to the image feature database to obtain a corresponding property judgment model, wherein the property judgment model is used for predicting the probability that the pulmonary mediastinal lymph node of the patient is malignant, namely the output of the property judgment model is the predicted malignant probability, the value range of the predicted malignant probability is [0,1], the more the predicted malignant probability output by the property judgment model is close to 0, the higher the possibility that the pulmonary mediastinal lymph node of the patient is benign is, the more the predicted malignant probability output by the property judgment model is close to 1, the higher the possibility that the pulmonary mediastinal lymph node of the patient is malignant is, and a doctor can use the predicted malignant probability as the reference of clinical diagnosis.
In a preferred embodiment of the present invention, the high-dimensional feature data comprises morphological feature data, and/or texture feature data, and/or grayscale feature data of the image of the lesion.
In a preferred embodiment of the present invention, the true malignancy probability is 0 or 1.
In a preferred embodiment of the present invention, the model training module 5 specifically includes:
a data dividing unit 51, configured to divide each focus data set in the image feature database into a training set and a test set;
the data training unit 52 is connected with the data dividing unit 51 and used for obtaining a property judgment model of the lung mediastinal lymph node through training according to each focus data set in the training set;
the model evaluation unit 53 is respectively connected with the data dividing unit 51 and the data training unit 52, and is used for inputting each focus data set in the test set into the property judgment model to obtain corresponding predicted malignancy probability, comparing each predicted malignancy probability with the corresponding real malignancy probability, and performing efficiency evaluation on the property judgment model according to the comparison result to obtain an efficiency evaluation result;
and the data output unit 54 is respectively connected with the data training unit 52 and the model evaluation unit 53 and is used for outputting the property judgment model and the performance evaluation result, and the doctor evaluates the accuracy of the predicted malignancy probability of the property judgment model according to the performance evaluation result and uses the evaluation as a clinical diagnosis reference.
Specifically, in the embodiment, the image feature database is divided into the training set and the test set, and the performance evaluation is performed on the property judgment model obtained by training the training set through the test set, so that the accuracy of the output predicted malignancy probability is provided while the predicted malignancy probability is provided for a doctor, a reference is further provided for clinical diagnosis of the doctor, and the reliability of the property judgment model is effectively improved.
A method for judging the properties of a pulmonary mediastinal lymph node, which is applied to any one of the above systems for judging the properties of a pulmonary mediastinal lymph node, as shown in fig. 2, the method for judging the properties specifically comprises the following steps:
step S1, collecting lung tomography images of a plurality of patients, and carrying out image segmentation on each lung tomography image to obtain focus position images of lung mediastinal lymph nodes corresponding to each lung tomography image;
step S2, respectively extracting the features of each focus position image to obtain high-dimensional feature data of each focus position image;
step S3, acquiring the real malignancy probability of the lung mediastinal lymph node corresponding to the focus position image aiming at each focus position image, and adding the focus position image, the high-dimensional characteristic data and the real malignancy probability into a focus data set;
step S4, constructing an image characteristic database according to each focus data set;
step S5, training according to the image feature database to obtain a property judgment model of the lung mediastinal lymph node;
step S6, inputting the lung tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignancy probability of the lung mediastinal lymph node of the patient to be predicted, so as to be used as reference for clinical diagnosis of doctors.
In a preferred embodiment of the present invention, in step S2, the high-dimensional feature data includes morphological feature data, texture feature data, and/or grayscale feature data of the lesion site image.
In the preferred embodiment of the present invention, in step S3, the true malignancy probability is 0 or 1.
In a preferred embodiment of the present invention, as shown in fig. 3, step S5 specifically includes:
step S51, dividing each focus data set in the image feature database into a training set and a testing set;
step S52, training according to each focus data set in the training set to obtain a property judgment model of the lung mediastinal lymph node;
step S53, inputting each focus data set in the test set into the property judgment model to obtain corresponding predicted malignancy probability, comparing each predicted malignancy probability with the corresponding real malignancy probability, and performing effectiveness evaluation on the property judgment model according to the comparison result to obtain an effectiveness evaluation result;
and step S54, outputting the property judgment model and the performance evaluation result, and evaluating the accuracy of the predicted malignancy probability of the property judgment model by the doctor according to the performance evaluation result and using the result as a clinical diagnosis reference.
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 detail may be made therein without departing from the spirit and scope of the invention.
Claims (8)
1. A system for determining properties of a mediastinal lymph node of a lung, comprising:
the image segmentation module is used for acquiring lung tomography images of a plurality of patients and performing image segmentation on each lung tomography image to obtain a lesion site image of the lung mediastinal lymph node corresponding to each lung tomography image;
the characteristic extraction module is connected with the image segmentation module and used for respectively extracting the characteristics of the images of the focus parts to obtain high-dimensional characteristic data of the images of the focus parts;
the data integration module is respectively connected with the image segmentation module and the feature extraction module and is used for acquiring the real malignant probability of the lung mediastinal lymph nodes corresponding to the focus position image aiming at each focus position image and adding the focus position image, the high-dimensional feature data and the real malignant probability into a focus data set;
the database construction module is connected with the data integration module and used for constructing an image characteristic database according to each focus data set;
the model training module is connected with the database construction module and used for obtaining a property judgment model of the lung mediastinal lymph node according to the image feature database;
and the model prediction module is connected with the model training module and used for inputting the pulmonary tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignant probability of the pulmonary mediastinal lymph node of the patient to be predicted so as to be referred to clinical diagnosis by a doctor.
2. The system for determining the nature of a mediastinal lymph node of a lung according to claim 1, wherein the high-dimensional feature data comprises morphological feature data, and/or texture feature data, and/or grayscale feature data of the image of the focal region.
3. The system for determining the nature of a regional mediastinal lymph node according to claim 1, wherein the true malignancy probability is 0 or 1.
4. The system for determining the nature of a mediastinal lymph node according to claim 1, wherein the model training module specifically comprises:
the data dividing unit is used for dividing each focus data set in the image feature database into a training set and a testing set;
the data training unit is connected with the data dividing unit and used for training according to each focus data set in the training set to obtain a property judgment model of the lung mediastinal lymph node;
the model evaluation unit is respectively connected with the data dividing unit and the data training unit and used for inputting each focus data set in the test set into the property judgment model to obtain corresponding predicted malignancy probability, comparing each predicted malignancy probability with the corresponding real malignancy probability, and carrying out efficiency evaluation on the property judgment model according to the comparison result to obtain an efficiency evaluation result;
and the data output unit is respectively connected with the data training unit and the model evaluation unit and is used for outputting the property judgment model and the performance evaluation result, and the doctor evaluates the accuracy of the predicted malignancy probability of the property judgment model according to the performance evaluation result and uses the evaluation result as a clinical diagnosis reference.
5. A method for determining a property of a pulmonary mediastinal lymph node, which is applied to the system for determining a property of a pulmonary mediastinal lymph node according to any one of claims 1 to 4, the method specifically comprising the steps of:
step S1, collecting lung tomography images of a plurality of patients, and carrying out image segmentation on the lung tomography images to obtain focus position images of lung mediastinal lymph nodes corresponding to the lung tomography images;
step S2, respectively carrying out feature extraction on each focus position image to obtain high-dimensional feature data of each focus position image;
step S3, for each focus position image, obtaining a true malignancy probability of the lung mediastinal lymph node corresponding to the focus position image, and adding the focus position image, the high-dimensional feature data and the true malignancy probability into a focus data set;
step S4, constructing an image characteristic database according to each focus data set;
step S5, training according to the image feature database to obtain a property judgment model of the lung mediastinal lymph node;
step S6, inputting the lung tomography image of the patient to be predicted into the property judgment model to obtain the predicted malignancy probability of the lung mediastinal lymph node of the patient to be predicted, so that doctors can make reference for clinical diagnosis.
6. The method for determining the nature of a pulmonary mediastinal lymph node according to claim 5, wherein the high-dimensional feature data includes morphological feature data, texture feature data, and/or grayscale feature data of the lesion site image in step S2.
7. The method for determining the property of a pulmonary mediastinal lymph node according to claim 5, wherein in the step S3, the true malignancy probability is 0 or 1.
8. The method for determining the property of a pulmonary mediastinal lymph node according to claim 5, wherein the step S5 specifically comprises:
step S51, dividing each focus data set in the image feature database into a training set and a testing set;
step S52, training according to each focus data set in the training set to obtain a property judgment model of the lung mediastinal lymph node;
step S53, inputting each focus data set in the test set into the property judgment model to obtain corresponding predicted malignancy probability, comparing each predicted malignancy probability with the corresponding real malignancy probability, and performing efficacy evaluation on the property judgment model according to the comparison result to obtain an efficacy evaluation result;
and step S54, outputting the property judgment model and the performance evaluation result, and evaluating the accuracy of the predicted malignancy probability of the property judgment model by the doctor according to the performance evaluation result and using the result as a clinical diagnosis reference.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010148269.XA CN111265234A (en) | 2020-03-05 | 2020-03-05 | Method and system for judging properties of lung mediastinal lymph nodes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010148269.XA CN111265234A (en) | 2020-03-05 | 2020-03-05 | Method and system for judging properties of lung mediastinal lymph nodes |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111265234A true CN111265234A (en) | 2020-06-12 |
Family
ID=70991476
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010148269.XA Pending CN111265234A (en) | 2020-03-05 | 2020-03-05 | Method and system for judging properties of lung mediastinal lymph nodes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111265234A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111985536A (en) * | 2020-07-17 | 2020-11-24 | 万达信息股份有限公司 | Gastroscope pathological image classification method based on weak supervised learning |
| CN113261985A (en) * | 2021-06-29 | 2021-08-17 | 上海长征医院 | Medical image processing method, device, medium and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040101181A1 (en) * | 2002-07-12 | 2004-05-27 | University Of Chicago | Automated method and system for computerized image analysis prognosis |
| CN106355023A (en) * | 2016-08-31 | 2017-01-25 | 北京数字精准医疗科技有限公司 | Open quantitative analysis method and system based on medical image |
| CN107280697A (en) * | 2017-05-15 | 2017-10-24 | 北京市计算中心 | Lung neoplasm grading determination method and system based on deep learning and data fusion |
| CN109117871A (en) * | 2018-07-20 | 2019-01-01 | 上海市肺科医院 | A kind of system and method that the pernicious probability of pulmonary lesions is obtained by deep learning |
-
2020
- 2020-03-05 CN CN202010148269.XA patent/CN111265234A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040101181A1 (en) * | 2002-07-12 | 2004-05-27 | University Of Chicago | Automated method and system for computerized image analysis prognosis |
| CN106355023A (en) * | 2016-08-31 | 2017-01-25 | 北京数字精准医疗科技有限公司 | Open quantitative analysis method and system based on medical image |
| CN107280697A (en) * | 2017-05-15 | 2017-10-24 | 北京市计算中心 | Lung neoplasm grading determination method and system based on deep learning and data fusion |
| CN109117871A (en) * | 2018-07-20 | 2019-01-01 | 上海市肺科医院 | A kind of system and method that the pernicious probability of pulmonary lesions is obtained by deep learning |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111985536A (en) * | 2020-07-17 | 2020-11-24 | 万达信息股份有限公司 | Gastroscope pathological image classification method based on weak supervised learning |
| CN111985536B (en) * | 2020-07-17 | 2024-02-09 | 万达信息股份有限公司 | Based on weak supervised learning gastroscopic pathology image Classification method |
| CN113261985A (en) * | 2021-06-29 | 2021-08-17 | 上海长征医院 | Medical image processing method, device, medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220319008A1 (en) | Automated tumor identification and segmentation with medical images | |
| JP5159242B2 (en) | Diagnosis support device, diagnosis support device control method, and program thereof | |
| Ali et al. | A systematic review of automated melanoma detection in dermatoscopic images and its ground truth data | |
| JP6807820B2 (en) | Image search device, method and program | |
| US7672491B2 (en) | Systems and methods providing automated decision support and medical imaging | |
| JP5814504B2 (en) | Medical image automatic segmentation system, apparatus and processor using statistical model | |
| US11915822B2 (en) | Medical image reading assistant apparatus and method for adjusting threshold of diagnostic assistant information based on follow-up examination | |
| JP2009516551A (en) | Quantitative and qualitative computer-aided analysis method and system for medical images | |
| JP5456132B2 (en) | Diagnosis support device, diagnosis support device control method, and program thereof | |
| CN111265234A (en) | Method and system for judging properties of lung mediastinal lymph nodes | |
| Paliwal et al. | A Comprehensive Analysis of Identifying Lung Cancer via Different Machine Learning Approach | |
| CN118435235A (en) | System and method for classifying lesions | |
| Bhushan | Liver cancer detection using hybrid approach-based convolutional neural network (HABCNN) | |
| Allapakam et al. | A hybrid feature pyramid network and Efficient Net-B0-based GIST detection and segmentation from fused CT-PET image | |
| Widodo et al. | Texture feature extraction to improve accuracy of malignant and benign cancer detection on ct-scan images | |
| JP2023541646A (en) | Biomarker estimation method | |
| CN118193770B (en) | Medical image retrieval method and system based on deep learning | |
| CN112862822B (en) | Ultrasonic breast tumor detection and classification method, device and medium | |
| Thool et al. | Artificial Intelligence in Medical Imaging Data Analytics using CT Images | |
| Khattar et al. | A Review on Preprocessing, Segmentation and Classification Techniques for Detection of Skin Cancer | |
| Sureka | Artificial Intelligence in Medical Imaging | |
| Babu et al. | Diagnosis of Lung Cancer Detection Using Various Deep Learning Models: A Survey | |
| Shakya et al. | An Adaptive Deep Learning Technique for Breast Cancer Diagnosis Based on Dataset | |
| CN118471443A (en) | Image histology data processing system for gastric cancer T stage diagnosis | |
| WO2023233031A1 (en) | Methods of radiomics features extraction |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200612 |
|
| RJ01 | Rejection of invention patent application after publication |