CN110084800B - Lung metastasis prediction method for patients with limb soft tissue sarcoma - Google Patents
Lung metastasis prediction method for patients with limb soft tissue sarcoma Download PDFInfo
- Publication number
- CN110084800B CN110084800B CN201910350067.0A CN201910350067A CN110084800B CN 110084800 B CN110084800 B CN 110084800B CN 201910350067 A CN201910350067 A CN 201910350067A CN 110084800 B CN110084800 B CN 110084800B
- Authority
- CN
- China
- Prior art keywords
- data
- tumor
- suv
- region
- lung metastasis
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- 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
- G06T2207/10104—Positron emission tomography [PET]
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Radiology & Medical Imaging (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Nuclear Medicine (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a lung metastasis prediction method for patients with limb soft tissue sarcoma, which comprises the following steps: s1, extracting the characteristics of all the tested objects by using the standard shooting value of the PET image, wherein the characteristics comprise SUV characteristics and texture characteristics; s2, extracting the characteristics of all tested clinical information by using a one-hot coding method; s3, sorting all the features by using a random forest algorithm, fusing the features with high contribution, and constructing a lung metastasis prediction model of the patient with the four-limb soft tissue sarcomas by using a BP neural network. The invention can more accurately predict the lung metastasis condition of the patients with the limb soft tissue sarcoma by using fewer patient characteristics.
Description
Technical Field
The invention relates to a lung metastasis prediction system and method for patients with limb soft tissue sarcoma, in particular to a lung metastasis prediction system and method for patients with limb soft tissue sarcoma by fusing PET images and clinical characteristics.
Background
Sarcomas are a highly heterogeneous group of tumors, classified according to the similar adult tissue types in histogenesis. It is characterized by invasive or destructive growth, recurrence and distant metastasis. As one of the sarcomas, soft tissue sarcomas (sts) can occur anywhere in the body, 59% of which originate in the limbs. Unfortunately, 10% -20% of patients with sarcoma or STS have distant metastases at the time of diagnosis. The metastasis rate during follow-up is about 30% to 40%, with lung metastasis accounting for about 90%. In addition, there is a great shortage of knowledge of prognostic factors for resection of pulmonary metastasis, and the recurrence rate after resection is high. Therefore, early screening and prediction of lung metastasis can help STS patients find corresponding treatment measures at an early stage, and the survival rate of the patients is improved.
The most common method of assessing the risk of lung metastasis is to study tumor heterogeneity from histopathological samples, while the biological relationship between different clonal subgroups or the microenvironment of the clones and the solid tumor, such as STS, is not clear, so that the obtained sample information is also affected by the sampled area and the solid cancer is spatially and temporally heterogeneous. Therefore, it is difficult to study the heterogeneity of tumors from the molecular level. Extracting a large number of features from medical images can solve this problem because radiology has the ability to capture intratumoral heterogeneity in a non-invasive manner. Features obtained from only a single STS image are limited and more other modality data, such as clinical data, may be ignored.
Therefore, the existing method for constructing the lung metastasis prediction model of the patient with the four-limb soft tissue sarcomas by using the image data is yet to be further developed and improved, and a more complete technical scheme needs to be provided on the basis of more in-depth research.
Disclosure of Invention
The invention aims to provide a lung metastasis prediction method for patients with limb soft tissue sarcoma, which integrates PET images and clinical characteristics, utilizes a random forest bag-out method to carry out characteristic sorting and select characteristics with high contribution degree, and constructs a lung metastasis prediction model of the limb soft tissue sarcoma through a BP neural network.
In order to achieve the purpose, the invention is realized by the following technical scheme:
a lung metastasis prediction method for a patient with limb soft tissue sarcoma is characterized by comprising the following steps:
s1, acquiring clinical data and PET image data of the patient with the sarcoma of the soft tissues of the limbs;
s2, fusing the acquired PET image data with a pre-calibrated tumor region to obtain an interested region, simultaneously calculating an SUV (motion-related vision) and merging and sorting clinical data, wherein the SUV is a standard uptake value;
s3, extracting the characteristics of the interested area and extracting the text characteristics of the clinical data by using a one-hot coding method;
s4, combining all the characteristics, and sorting the characteristic contribution degrees of all the characteristics by using a random forest algorithm and a double-sample T test algorithm;
s5, selecting the characteristics with high contribution degree, and constructing a lung metastasis prediction model of the patient with the sarcoma of the soft tissues of the limbs by utilizing a back propagation neural network.
The step S2 includes the following steps:
s2.1, extracting tumor region data from the collected data, specifically: for each tested object, positioning a tumor voxel region in the raw data of the PET image of the patient with the four-limb soft tissue sarcoma by using the tumor region outlined by a professional doctor, marking the tumor voxel region as an interested region, and then extracting the interested region of each tested object;
s2.2, calculating SUV in the interested region, and simultaneously merging and sorting all clinical data.
The step S3 includes:
s3.1, extracting SUV features of the PET image, wherein the SUV features comprise SUVmax, SUVpeak, SUVmean, AUC-CSH and PercentInactive, and the SUVmax represents the maximum SUV value of a tumor region; SUVpeak represents the average between the tumor region and its neighbors, taking the maximum value; SUVmean represents the mean SUV value for the tumor region; AUC-CSH represents the area under the cumulative standard uptake value volume histogram curve, describing the percentage of total tumor volume that exceeds the maximum SUV threshold; PercentInactive represents the percentage of inactive tumor area;
s3.2, performing text feature extraction on clinical data: the method comprises the following steps of utilizing a one-hot coding method to extract features of clinical text information, wherein the clinical text information comprises the following steps: age, sex, survival status, level of tumor differentiation, tumor type, tumor location, or treatment modality.
The step S4 includes:
s4.1, fusing all the characteristics and combining the characteristics to obtain a characteristic matrix, wherein the row samples are listed as all the characteristics, the last column represents a label for judging whether each tested lung is transferred, the lung is transferred to be 1, the non-lung is transferred to be 0, based on the characteristic matrix and each tested label, all the tested subjects are randomly divided into N groups by a resampling method, one part of data in each group is divided into in-bag data, namely a training set, and the rest part of data is used as out-bag data, namely test data.
S4.2, the N groups of data correspond to N decision trees in the random forest, and for each decision tree, the error of the data outside the bag is calculated by using the corresponding data outside the bag and is recorded as err 1.
S4.3, noise interference is added to the characteristics M of all samples of the data outside the bag at random, the error of the data outside the bag of the decision tree is calculated again and is marked as err2, and therefore, the importance of the characteristics M can be calculated as IM=∑(err1-err2) N is represented by formula IMThe larger the value, the higher the importance of the feature M and the higher the contribution to the predictive model.
The step S5 includes:
s5.1, randomly dividing all the tested objects into three groups, including a training set, a testing set and a verification set.
S5.2, constructing a BP neural network model based on the features with high contribution degree selected by the random forest, setting the neuron number of an input layer as a feature number, setting the neuron number of a hidden layer, setting the neuron number of an output layer as a label type number, and setting the highest iteration number, the minimum gradient, the initial parameter randomness and an activation function.
S5.3, randomly disordering the sequence of the samples, repeating the operation of the step S5.1 and the operation of the step S5.2 for N times, and taking the average value of the N times as a model performance result to avoid local optimization, wherein the model with the highest accuracy, sensitivity and characteristic of the training set, the verification set and the test set is the final prediction model, and the model with the best performance is selected as the final prediction model.
Compared with the prior art, the invention has the following advantages:
1. a feature fusion method based on PET images and clinical information is provided, and is used for obtaining features of four-limb soft tissue sarcoma lung metastasis prediction.
2. A method of machine learning is used, including random forest and BP neural networks, to select valid features by eliminating irrelevant or redundant features, and then to construct a predictive model of lung metastasis. The predictive power of the proposed fused feature model is significantly better than that of the model using image or clinical features alone, including accuracy, specificity, or sensitivity.
Drawings
FIG. 1 is a flow chart of a method for predicting lung metastasis in patients with soft tissue sarcoma of the extremities according to the present invention.
Detailed Description
The present invention will now be further described by way of the following detailed description of a preferred embodiment thereof, taken in conjunction with the accompanying drawings.
As shown in fig. 1, a method for predicting lung metastasis in a patient with soft tissue sarcoma of limbs comprises:
step 1: acquiring clinical data and PET image data of patients with limb soft tissue sarcoma, wherein the PET image data is in DICOM format and comprises 51 data to be tested, wherein
Step 2: fusing the acquired image data with a pre-calibrated tumor region to obtain a region of interest (ROI), calculating a standard uptake value (SUV value), and merging and sorting clinical data;
in a specific embodiment, the step 2 comprises:
step 2.1, extracting tumor region data from the acquired data, specifically including the following steps of positioning a tumor voxel region in original DICOM data of a PET image of a patient with limb soft tissue sarcoma by using a tumor region outlined by a professional doctor for each tested object, marking the tumor voxel region as ROI, and then extracting the ROI of each tested object;
step 2.2, calculating a Standard Uptake Value (SUV) of the ROI area, wherein the specific method comprises the following steps: the SUV value for each tumor section tested was calculated one by one using the formula SUV-radioactive concentration/injected dose of the lesion-body weight. And simultaneously merging and sorting all clinical data.
And step 3: based on the preprocessed data, performing feature extraction on the ROI, wherein the feature extraction comprises SUV features and textural features, and performing text feature extraction on clinical data by using a one-hot coding method;
in a specific embodiment, the step 3 comprises:
and 3.1, extracting the SUV features of the PET image, wherein the specific method is as follows, and the SUV features comprise 5, namely SUVmax, SUVpeak, SUVmean, AUC-CSH and PercentInactive. Wherein SUVmax represents the maximum SUV value for the tumor region; SUVpeak represents the average between the tumor region and its neighbors, taking the maximum value; SUVmean represents the mean SUV value for the tumor region; AUC-CSH represents the area under the cumulative SUV volume histogram curve, describing the percentage of total tumor volume that exceeds the maximum SUV threshold; PercentInactive represents the percentage of inactive tumor area.
And 3.2, extracting the clinical characteristics of the patients with the sarcoma of the soft tissues of the limbs, wherein the specific method comprises the following step of extracting the characteristics of clinical text information by using a one-hot coding method, for example, the sex comprises male and female, the male code is 01, the female code is 10, and other text characteristics are selected according to the characteristics of the method, including age, sex, survival state, tumor differentiation level, tumor type, tumor position and treatment mode.
In a specific embodiment, the step 4 includes:
and 4, step 4: combining all the features, and performing feature contribution degree sequencing on all the features by using a random forest algorithm and a double-sample T test algorithm, wherein the combination of all the features comprises the following steps: SUV features, texture features, and text features;
and 4.1, fusing all the characteristics and combining the characteristics to obtain a characteristic matrix, wherein the behavior samples are listed as all the characteristics, the last column represents a label of whether each tested lung is metastatic or not, the lung metastasis is 1, the non-lung metastasis is 0, and all the tested lungs are randomly divided into N groups by using a Bootstrap resampling method based on the characteristic matrix and each tested label, wherein 2/3 data in each group are used as in-bag data, namely training sets, and 1/3 data are used as out-of-bag data, namely test data.
And 4.2, the N groups of data correspond to N decision trees in the random forest, and for each decision tree, calculating the error of the data outside the bag by using the corresponding data outside the bag, and recording the error as err 1.
Step 4.3, noise is added to the characteristics M of all samples of the data outside the bag randomlyAcoustic interference, the error of the data outside the bag of the decision tree is calculated again, denoted as err2, and therefore the importance to the feature M can be calculated as IM=∑(err1-err2) N is represented by formula IMThe larger the value, the higher the importance of the feature M and the higher the contribution to the predictive model.
And 5: and selecting the characteristic with high contribution degree, and constructing a lung metastasis prediction model of the patient with the sarcoma of the soft tissues of the limbs by using the BP neural network.
In a particular embodiment, this step 5 comprises:
and 5.1, randomly dividing all the tested objects into three groups including a training set, a testing set and a verification set.
And 5.2, constructing a BP neural network model based on the features with high contribution degree selected by the random forest, setting the number of neurons in an input layer as the feature number, the number of neurons in a hidden layer as 10, the number of neurons in an output layer as 2, setting the highest iteration number as 1000, setting the minimum gradient as 0.001, setting the initial parameters as random, and setting the activation function as a sigmod function.
And 5.3, randomly disordering the sequence of the samples, repeating the operation of the step 5.1 and the operation of the step 5.2 for 100 times, and taking the average value of the 100 times as a model performance result to avoid local optimization, wherein the model with the highest accuracy, sensitivity and characteristic of the training set, the verification set and the test set is the final prediction model, and the model with the best performance is selected as the final prediction model.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.
Claims (2)
1. A lung metastasis prediction method for patients with limb soft tissue sarcoma is characterized by comprising the following steps:
s1, acquiring clinical data and PET image data of the patient with the sarcoma of the soft tissues of the limbs;
s2, fusing the acquired PET image data with a pre-calibrated tumor region to obtain an interested region, simultaneously calculating an SUV (motion-related vision) and merging and sorting clinical data, wherein the SUV is a standard uptake value;
s3, extracting the characteristics of the interested area and extracting the text characteristics of the clinical data by using a one-hot coding method;
s4, combining all the characteristics, and sorting the characteristic contribution degrees of all the characteristics by using a random forest algorithm and a double-sample T test algorithm;
s5, selecting the characteristics with high contribution degree, and constructing a lung metastasis prediction model of the patient with the sarcoma of the soft tissues of the limbs by utilizing a back propagation neural network;
the step S2 includes the following steps:
s2.1, extracting tumor region data from the collected data, specifically: for each tested object, positioning a tumor voxel region in the raw data of the PET image of the patient with the four-limb soft tissue sarcoma by using the tumor region outlined by a professional doctor, marking the tumor voxel region as an interested region, and then extracting the interested region of each tested object;
s2.2, calculating the SUV in the interested region, and simultaneously merging and sorting all clinical data;
the step S3 includes:
s3.1, extracting SUV features of the PET image, wherein the SUV features comprise SUVmax, SUVpeak, SUVmean, AUC-CSH and PercentInactive, and the SUVmax represents the maximum SUV value of a tumor region; SUVpeak represents the average between the tumor region and its neighbors, taking the maximum value; SUVmean represents the mean SUV value for the tumor region; AUC-CSH represents the area under the cumulative standard uptake value volume histogram curve, describing the percentage of total tumor volume that exceeds the maximum SUV threshold; PercentInactive represents the percentage of inactive tumor area;
s3.2, performing text feature extraction on clinical data: the method comprises the following steps of utilizing a one-hot coding method to extract features of clinical text information, wherein the clinical text information comprises the following steps: one or more of age, sex, survival status, level of tumor differentiation, tumor type, tumor location, or treatment modality; the step S4 includes:
s4.1, fusing all the characteristics and combining the characteristics to obtain a characteristic matrix, wherein the row samples are listed as all the characteristics, the last column of the characteristic matrix is a label for indicating whether each tested lung is metastatic or not, the lung metastasis is 1, the non-lung metastasis is 0, based on the characteristic matrix and each tested label, all the tested subjects are randomly divided into N groups by a resampling method, one part of data in each group is divided into in-bag data, namely a training set, and the rest part of data is used as out-bag data, namely test data;
s4.2, the N groups of data correspond to N decision trees in the random forest, and for each decision tree, the error of the data outside the bag is calculated by using the corresponding data outside the bag and is recorded as err1 ;
S4.3, randomly adding noise interference to the characteristics M of all samples of the data outside the bag, and calculating the error of the data outside the bag of the decision tree again and recording the error as err2Thus, the importance to feature M can be calculated as IM=∑(err1-err2) N is represented by formula IMThe larger the value, the higher the importance of the feature M and the higher the contribution to the predictive model.
2. The method for predicting lung metastasis in patients with soft tissue sarcoma of extremities of claim 1, wherein the step S5 comprises:
s5.1, randomly dividing all the tested objects into three groups, including a training set, a testing set and a verification set;
s5.2, constructing a BP neural network model based on the features with high contribution degree selected by the random forest, setting the neuron number of an input layer as a feature number, setting the neuron number of a hidden layer, setting the neuron number of an output layer as a label type number, and setting the highest iteration number, the minimum gradient, the initial parameter random and an activation function;
s5.3, randomly disordering the sequence of the samples, repeating the operation of the step S5.1 and the operation of the step S5.2 for N times, and taking the average value of the N times as a model performance result to avoid local optimization, wherein the model with the highest accuracy, sensitivity and characteristic of the training set, the verification set and the test set is the final prediction model, and the model with the best performance is selected as the final prediction model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910350067.0A CN110084800B (en) | 2019-04-28 | 2019-04-28 | Lung metastasis prediction method for patients with limb soft tissue sarcoma |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910350067.0A CN110084800B (en) | 2019-04-28 | 2019-04-28 | Lung metastasis prediction method for patients with limb soft tissue sarcoma |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110084800A CN110084800A (en) | 2019-08-02 |
CN110084800B true CN110084800B (en) | 2021-03-16 |
Family
ID=67417293
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910350067.0A Active CN110084800B (en) | 2019-04-28 | 2019-04-28 | Lung metastasis prediction method for patients with limb soft tissue sarcoma |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110084800B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110516759B (en) * | 2019-09-02 | 2022-09-30 | 河南师范大学 | Soft tissue sarcoma transfer risk prediction system based on machine learning |
CN115171905B (en) * | 2022-06-20 | 2023-04-07 | 复旦大学 | Tumor patient similarity calculation method based on one-hot coding unsupervised clustering |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB0913314D0 (en) * | 2009-07-31 | 2009-09-02 | Siemens Medical Solutions | Facilitated percist evaluation |
US8824752B1 (en) * | 2013-03-15 | 2014-09-02 | Heartflow, Inc. | Methods and systems for assessing image quality in modeling of patient anatomic or blood flow characteristics |
CN104463840A (en) * | 2014-09-29 | 2015-03-25 | 北京理工大学 | Fever to-be-checked computer aided diagnosis method based on PET/CT images |
CN105869137B (en) * | 2015-01-22 | 2019-02-19 | 山东师范大学 | A kind of extracting method of the tumour textural characteristics based on multi-scale technique |
CN105427325A (en) * | 2015-12-07 | 2016-03-23 | 苏州大学 | Automatic lung tumour segmentation method based on random forest and monotonically decreasing function |
CN106251313B (en) * | 2016-08-15 | 2020-06-26 | 上海联影医疗科技有限公司 | Medical imaging method and system |
-
2019
- 2019-04-28 CN CN201910350067.0A patent/CN110084800B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN110084800A (en) | 2019-08-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11615879B2 (en) | System and method for automated labeling and annotating unstructured medical datasets | |
Sahu et al. | FINE_DENSEIGANET: Automatic medical image classification in chest CT scan using Hybrid Deep Learning Framework | |
US20220301714A1 (en) | Method for predicting lung cancer development based on artificial intelligence model, and analysis device therefor | |
CN111445946B (en) | Calculation method for calculating lung cancer genotyping by using PET/CT (positron emission tomography/computed tomography) images | |
CN110443268B (en) | Liver CT image benign and malignant classification method based on deep learning | |
CN110246577B (en) | Method for assisting gestational diabetes genetic risk prediction based on artificial intelligence | |
WO2024065987A1 (en) | Lung cancer prognosis prediction system based on multi-omics of radiomics, pathomics and genomics | |
CN112561869B (en) | Pancreatic neuroendocrine tumor postoperative recurrence risk prediction method | |
CN110084800B (en) | Lung metastasis prediction method for patients with limb soft tissue sarcoma | |
CN112365980A (en) | Brain tumor multi-target point auxiliary diagnosis and prospective treatment evolution visualization method and system | |
JP7197795B2 (en) | Machine learning program, machine learning method and machine learning apparatus | |
CN115019405A (en) | Multi-modal fusion-based tumor classification method and system | |
Aston et al. | Deep learning applied to attractor images derived from ECG signals for detection of genetic mutation | |
CN115896242A (en) | Intelligent cancer screening model and method based on peripheral blood immune characteristics | |
CN117116477A (en) | Construction method and system of prostate cancer disease risk prediction model based on random forest and XGBoost | |
CN108597615A (en) | A kind of screening reference method of Patients with Mild Cognitive Impairment dementia conversion | |
Hu et al. | A Hybrid Convolutional Neural Network Model Based on Different Evolution for Medical Image Classification. | |
Kolli et al. | Efficient development of supervised learning algorithm for kidney stone prediction | |
CN117079801B (en) | Colorectal cancer prognosis risk prediction system | |
Charrière et al. | A comparative study of two automated solutions for cross‐sectional skeletal muscle measurement from abdominal computed tomography images | |
CN110992312B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
CN116504406A (en) | Method and system for constructing lung cancer postoperative risk model based on image combination pathology | |
US20110071967A1 (en) | Automatic Labeler Assignment | |
CN101466302B (en) | A method to individualize kinetic parameter estimation for dynamic molecular imaging procedures | |
CN113362927A (en) | Squamous esophageal cancer chemoradiotherapy effect prediction method based on deep learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |