CN109583447A - A kind of image group credit analysis model building method and analysis method - Google Patents
A kind of image group credit analysis model building method and analysis method Download PDFInfo
- Publication number
- CN109583447A CN109583447A CN201811512310.6A CN201811512310A CN109583447A CN 109583447 A CN109583447 A CN 109583447A CN 201811512310 A CN201811512310 A CN 201811512310A CN 109583447 A CN109583447 A CN 109583447A
- Authority
- CN
- China
- Prior art keywords
- image
- model
- image group
- treatment
- feature
- 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
Classifications
-
- 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
- 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
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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]
-
- 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)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention provides one kind, the invention discloses a kind of image group credits to analyse model building method, comprising: obtains the image data of different types of patient at pre-treatment and after treatment;The interested area division from image data;Image feature extraction is carried out for the area-of-interest in the image data before treatment and after treatment respectively;Image feature based on extraction establishes image group analysis model.The present invention is included in the image group model of the CT data mining of pretreatment and post-processing by noninvasive mode, wieldy nomogram can be by the benefit degree in a manner of personalized prediction Patients with Non-small-cell Lung difference clinical treatment, to provide an effective tool for clinical decision.
Description
Technical field
This application involves biomedical engineering technology fields, and in particular, to a kind of image group credit analysis model construction side
Method and analysis method.
Background technique
First carrying out new chemoradiation therapy, the comprehensive therapeutic plans such as row radical surgery are a kind of basic therapeutic strategies again, but
It is for different types of patient, the effect of the strategy is that there are great differences.With non-small cell lung cancer (non-small
Cell lung cancer, NSCLC) for, research confirms that in terms of middle position progression free survival phase, visiting before operation has obviously
Advantage, and Distant metastasis rates are also below independent operation group;For I phase phase-II (N0-1) patient of no mediastinal lymph node metastasis, art
The existence benefit of preceding chemicotherapy becomes apparent from;And for IIIa phase (N2) patient, it does not have a clear superiority.
However, identifying different types of patient before treatment there is no correlation technique at present.
Summary of the invention
In order to solve the above-mentioned technical problem, the purpose of the present invention is to provide a kind of building sides of image group analysis model
Method and image group analysis method, can identify different types of patient before the treatment.
According to the first aspect of the invention, the following technical solution is proposed:
Model building method is analysed in a kind of image group credit according to the present invention characterized by comprising
S1, the image data of different types of patient at pre-treatment and after treatment is obtained;
S2, the interested area division from image data;
S3, the area-of-interest being directed in the image data before treatment and after treatment respectively carry out image feature extraction;
S4, the image feature based on extraction establish image group analysis model.
Further, model building method is analysed in image group credit according to the present invention, which is characterized in that further packet
It includes:
The treatment includes radiotherapy and/or chemotherapy.
Further, model building method is analysed in image group credit according to the present invention, and the image data includes CT
Image.
Model building method is analysed in image group credit according to the present invention, and the area-of-interest is tumor region.
Further, model building method is analysed in image group credit according to the present invention, and the image feature extracts packet
It includes:
It is examined using two sample T, and image spy is extracted using minimum absolute retract algorithm and selection operator regression algorithm
Sign.
Further, model building method is analysed in image group credit according to the present invention, is included the following steps:
Firstly, selecting complete pathology alleviation group in main queue and being non-fully based on univariate statistics between pathology alleviation group
The best features of inspection;
Secondly, extracting image feature using minimum absolute retract algorithm, will have minimum absolute retract and selection operator to punish
The canonical multipie logistic regression penalized is applied to the data of main queue, wherein the Result for Combinations of selected feature is each
From coefficient, a model is set and is used to estimate the chemicotherapy based on radiographic feature as a result, wherein the model is according to following
Mode is defined:
Wherein, indicate that patient is in complete pathology alleviation group when the value of model y is 1, the value of model y indicates patient when being 0
In non-fully pathology alleviation group;D is indicated with total number of variable in the model;xj(j=1,2 ... d) indicates variable;βj(j
=0,1,2 ... d) indicates model parameter, and ε indicates error term,
Again, using operator regression algorithm extract image feature, specifically, with the parameter of regularized regression estimation models,
Feature selecting, and can be carried out simultaneously by assigning many parameters with zero:
Wherein γiIndicate the result of patient i;N indicates number of patients;S is Sigmoid function;xijRepresent i-th of patient's
J-th of feature;λ indicates regularization parameter, wherein Sigmoid function is defined as:
Using what is punished with selection operator LASSOBy the way that some parameter betas are arrangedjIt is sparse to guide for 0
Model, then select to contribute model biggish feature come using.
Further, model building method, the image based on extraction are analysed in image group credit according to the present invention
Feature establishes image group analysis model
Image group analysis model is established using Multiple Logistic regression analysis.
According to a kind of this hair invention image group analysis method characterized by comprising
Image data before obtaining patient's treatment;
The interested area division from the image data;
Image feature extraction is carried out for the area-of-interest;
Image feature based on extraction and the image group according to above-mentioned image group credit analysis model building method building
Analysis model classifies to patient.
The one or more technical solutions provided in the embodiment of the present invention, have at least the following technical effects or advantages:
Image group analysis model is established based on big data thought, can be classified before the treatment to different patients.?
On the basis of carrying out Multiple Logistic regression analysis to the above-mentioned clinical parameter of main queue, PCR is constructed with the sample filtered out
The image group model of detection, provides quantitative tool for clinic.
It is all assumed to be in addition, the image group model proposed according to the present invention shows detecting pCR than all patients
To realize the bigger advantage of the scheme without patient.The shadow of the CT data mining of pretreatment and post-processing is included in by noninvasive mode
As group model, wieldy nomogram can obtaining in a manner of personalized prediction Patients with Non-small-cell Lung difference clinical treatment
Beneficial degree, to provide an effective tool for clinical decision.
Detailed description of the invention
Fig. 1 is that model building method schematic diagram is analysed in image group credit of the invention;
Fig. 2 is the performance comparison figure of image group feature of the present invention and image group model.
Specific embodiment
Following for the above objects, features, and advantages that can be more clearly understood that the application, with reference to the accompanying drawing and have
The application is further described in detail in body embodiment.It should be noted that in the absence of conflict, the application's
Feature in embodiment and embodiment can be combined with each other.
Many details are explained in the following description in order to fully understand the application, still, the application may be used also
To be implemented using other than the one described here other modes, therefore, the protection scope of the application is not by described below
Specific embodiment limitation.
Fig. 1 shows image group credit analysis model building method of the invention, comprising:
S1, the image data of different types of patient at pre-treatment and after treatment is obtained;
Preferably, the treatment includes radiotherapy and/or chemotherapy, and the image data includes CT images data.
S2, the interested area division from image data;
For the pretherapy and post-treatment image data of different types of patient, region of interest ROI, the area-of-interest are selected
For example, tumor region.
S3, the area-of-interest being directed in the image data before treatment and after treatment respectively carry out image feature extraction;
The image feature extracts
It is examined using two sample T, and extracts image spy using minimum absolute retract algorithm and selection operator regression algorithm
Sign, it is therefore intended that in order to reduce overfitting or any kind of deviation.
Specifically, described examined using two sample T is mentioned using minimum absolute retract algorithm and selection operator regression algorithm
Image feature is taken, specific steps include:
(1) firstly, pCR (complete pathology alleviation) is organized in selection main queue and non-pCR (non-fully pathology alleviation) organizes it
Between based on univariate statistics examine (double sample t inspection) best features;
(2) secondly, extracting image feature using minimum absolute retract algorithm, specifically, will have minimum absolute retract and
The canonical multipie logistic regression of selection operator (LASSO) punishment is applied to the data of main queue, wherein by selected spy
The respective coefficient of the Result for Combinations of sign, be arranged a model be used to estimate the chemicotherapy based on radiographic feature as a result,
Wherein the model is defined in the following way:
Wherein, it indicates that patient is in pCR group when the value of model y is 1, indicates that patient is in non-pCR when the value of model y is 0
Group;D is indicated with total number of variable in the model;xj(j=1,2 ... d) indicates variable;βj(j=0,1,2 ... d) indicates mould
Shape parameter, ε indicate error term.
(3) again, image feature is extracted using selection operator regression algorithm, specifically, using regularized regression
The parameter, feature selecting (by assigning many parameters with zero) of (regularized regression) estimation models can be same
Shi Jinhang.The purpose of this method is to minimize cost function:
Wherein γiIndicate the result of patient i;N indicates number of patients;S is Sigmoid function;xijRepresent i-th of patient's
J-th of feature;λ indicates regularization parameter, wherein Sigmoid function is defined as:
With selection operator LASSO punishmentIt is applied, by the way that some parameter betas are arrangedjIt is dilute to guide for 0
Dredge model, then select to contribute model biggish feature come using.
Wherein, sample T described in step S13 is examined, and also known as Student T examines (Student's T test), mainly
Smaller (such as n < 30), the population standard deviation σ unknown normal distribution for sample content.
S4, the image feature based on extraction establish image group analysis model.
It includes: using Multiple Logistic regression analysis that the image feature based on extraction, which establishes image group analysis model,
Establish image group analysis model.
Further, after model foundation is analysed in image group credit, image group analysis model is assessed.Institute's commentary
Estimate includes: that repeatability assessment and/or Hosmer-Clemeshow are examined between observer in observer.
Further, a kind of image group analysis method of the invention specifically comprises the following steps:
S21, the image data before patient's treatment is obtained;
S22, the image feature based on extraction and the image group analysis model constructed according to above-mentioned method, to patient into
Row classification.
According to one embodiment of present invention, still with non-small cell lung cancer (non-small cell lung cancer,
NSCLC for), patient is divided into two types: one is reaching pCR (complete pathology alleviation), referred to as PCR group after the treatment,
Another kind is to be not up to pCR, referred to as non-pCR group after the treatment.The embodiment for establishing image group analysis model is as follows:
Obtain the CT images before different types of patient's chemicotherapy and after chemicotherapy.The CT images that patient per is CT all can
Storage in the database, the patient's CT images and the second quantity of the pCR group of the first quantity is therefrom chosen according to the effect after treatment
Non- pCR group patient's CT images.First quantity and the second quantity can be the same or different, and quantity is more, ultimately forms
Model differentiate when accuracy it is higher.In such an embodiment, it is preferred to first quantity and the second quantity both greater than or
Equal to 100.
Area-of-interest (the regions of interest, ROI) is created to each CT images.Area-of-interest makes
With strengthening imaging and nonreinforcement imaging data, including entire tumour and exclude the lobe of the lung.Before chemicotherapy, shown along imaging data is strengthened
Tumor's profiles draw ROI, string and burr comprising surrounding.Region of interest is placed in the nonreinforcement imaging area on each section CT
Domain.If strengthening imaging after chemicotherapy height suspected tumor signal still occurs, it is identical as before chemicotherapy that ROI delineates standard.If
Detected in tumor bed on strengthening imaging data it is low, mix intensity or any other improper lobe of the lung signal (abnormal letter
Number), then ROI is drawn with the profile in abnormal signal region.In the case where reinforcing imaging data no abnormal signal, by ROI
Strengthen the primary tumor bed region that imaging data determines before being placed in chemicotherapy.If noticing that height can on nonreinforcement imaging data
Doubtful tumor signal (high RST), then be placed in high RST overlying regions for ROI.In nonreinforcement imaging data compared with the normal lobe of the lung
In the case where high RST is not detected, ROI is placed in the primary tumor bed region of nonreinforcement imaging data determination before chemicotherapy.If
After chemicotherapy strengthen imaging data have no tumor signal, then according to pretreatment image corresponding tumor bed region delineate nonreinforcement at
As the ROI of data.
Each CT images Z value is normalized, to obtain the standardized normal distribution of image intensity.In order to reduce overfitting
Or any kind of deviation has used two feature selection steps in our Radiomics model.Firstly, selection is main
PCR (complete pathology alleviation) examines the best spy of (double sample t inspection) based on univariate statistics in queue between group and non-pCR group
Sign.Secondly, will have minimum absolute retract and the canonical multipie logistic regression of selection operator (LASSO) punishment to be applied to
The data of main queue.
3 groups of imagings spies are extracted from before standardized treatment and after treatment in the augmentation data of CT images and nonreinforcement data
Sign, and divide ROI:(i) 4 statistical natures, (ii) 43 voxel Strength co-mputation features, and (iii) 516 wavelet characters.1st
Group is made of the tumour strength characteristic quantified, first order statistic by all tumour intensity histogram calculation.2nd group comprising being based on
The textural characteristics (that is, the texture difference observed in gross tumor volume) of heterogeneous quantization in tumour;These features use two
Analytical calculation is tieed up, and all slices in three-dimensional nodule volume are averaged.3rd group is closed from the wavelet decomposition of original image
And calculated textural characteristics, to focus on the various dimensions in frequency and different characteristic orientations in gross tumor volume.Last
Group includes 563 features of every kind of mode (reinforcing and nonreinforcement) of each MRI scan, each a total of 2252 radiation of patient
Learn feature.
In an embodiment of the present invention, logistic regression analysis: age, property is carried out using following clinical information
Not, CEA after treatment, CA19-9 after treatment, histological grade swell before treatment and after treating with length of tumor after treatment before treatment
Tumor thickness, infringement distance is more than visceral pleura before treatment and after treatment, before treatment and outside visceral pleura after treatment and expanding tumor
The shortest distance between edge, lymph node is total (NLN) before the treatment of detection and after treatment, before the treatment of maximum lymph node and after treatment
Minor axis length (MALLLN) and radiation group feature.Using likelihood ratio test, advised using Akaike's Information Criterion as stopping
Then, using gradually selecting backward.On the basis of the above-mentioned clinical parameter to main queue carries out Multiple Logistic regression analysis,
With the image group model of the sample building PCR detection filtered out.
Further, in order to verify the effect of model, in application observation person between observer repeatability assessment and/or
Hosmer-Clemeshow is examined.If verification the verifying results are unable to meet demand, above step is continued to execute, increases more suffer from
Person's data construct image group model, until reaching standard.
It is shown according to the method for the present invention in test data wherein, the details of image group characteristic performance
As shown in Fig. 2 (performance comparison of image group feature and image group model).In Fig. 2, in verifying queue, radiation group
The AUC that feature generates is 0.9744 [95% credibility interval (CI), 0.9642-0.9756], classification accuracy 94.08%
(95% credibility interval, 93.19-94.79%);In verifying queue the AUC that generates be 0.9799 (95% credibility interval,
0.9780-0.9840), classification accuracy is 94.29% (95% credibility interval, 94.21-95.61%).Importantly, putting
Penetrate group feature realizes 86.96% PPV (95% credibility interval, 84.84-90.40%) in main queue, in verifying team
90.00% PPV (95% credibility interval, 89.60-99.40%) is realized in column.By Fig. 2 it can be found that in primary queue
The image group diversity of values of pCR and non-pCR patient are statistically significant (P < 0.01);Verifying queue be also so (P <
0.01) it, can satisfy verifying completely to require, i.e., verifying is recorded a demerit shows good consistency in main queue.
In this application, the description of term " one embodiment ", " some embodiments ", " specific embodiment " etc. means to combine
The particular features, structures, materials, or characteristics of embodiment or example description are contained at least one embodiment of the application or show
In example.In the present specification, schematic expression of the above terms are not necessarily referring to identical embodiment or example.Moreover,
The particular features, structures, materials, or characteristics of description can be in any one or more embodiment or examples in an appropriate manner
In conjunction with.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (8)
1. model building method is analysed in a kind of image group credit, which comprises the steps of:
S1, the image data of different types of patient at pre-treatment and after treatment is obtained;
S2, the interested area division from image data;
S3, the area-of-interest being directed in the image data before treatment and after treatment respectively carry out image feature extraction;
S4, the image feature based on extraction establish image group analysis model.
2. model building method is analysed in image group credit according to claim 1, which is characterized in that the treatment includes radiotherapy
And/or chemotherapy.
3. model building method is analysed in image group credit according to claim 2, which is characterized in that the image data includes
CT images.
4. model building method is analysed in image group credit according to claim 3, which is characterized in that the area-of-interest is
Tumor region.
5. model building method is analysed in image group credit according to claim 4, which is characterized in that the image feature extracts
Include:
It is examined using two sample T and extracts image feature with minimum absolute retract and selection operator regression algorithm.
6. model building method is analysed in image group credit described in one of -5 according to claim 1, which is characterized in that described to use two
A sample T is examined, and is extracted image feature using minimum absolute retract algorithm and selection operator regression algorithm and included the following steps:
Firstly, selecting complete pathology alleviation group in main queue and being non-fully based on univariate statistics inspection between pathology alleviation group
Best features;
Secondly, minimum absolute retract algorithm is used to extract image feature, will there is minimum absolute retract and selection operator to punish
Canonical multipie logistic regression is applied to the data of main queue, wherein the Result for Combinations of selected feature is respective
Coefficient is arranged a model and is used to estimate the chemicotherapy based on radiographic feature as a result, wherein the model is in the following way
It is defined:
Wherein, it indicates that patient is in complete pathology alleviation group when the value of model y is 1, indicates that patient is in when the value of model y is 0
Non-fully pathology alleviation group;D is indicated with total number of variable in the model;xj(j=1,2 ... d) indicates variable;βj(j=0,
1,2 ... d) indicates model parameter, and ε indicates error term,
Again, image feature is extracted using selection operator regression algorithm, is selected with the parameter of regularized regression estimation models, feature
It selects, and can be carried out simultaneously by assigning many parameters with zero:
Wherein γiIndicate the result of patient i;N indicates number of patients;S is Sigmoid function;xijRepresent the jth of i-th of patient
A feature;λ indicates regularization parameter, wherein Sigmoid function is defined as:
Using what is punished with selection operator LASSOSparse model is guided by the way that some parameter beta j of setting are 0,
Then select to contribute model biggish feature come using.
7. model building method is analysed in image group credit according to claim 6, the image feature based on extraction is established
Image group analysis model includes:
Image group analysis model is established using Multiple Logistic regression analysis.
8. a kind of image group analysis method characterized by comprising
Image data before obtaining patient's treatment;
The interested area division from the image data;
Image feature extraction is carried out for the area-of-interest;
Image feature based on extraction and according to the analysis model building method building of image group credit described in one of claim 1-7
Image group analysis model, classify to patient.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811512310.6A CN109583447A (en) | 2018-12-11 | 2018-12-11 | A kind of image group credit analysis model building method and analysis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811512310.6A CN109583447A (en) | 2018-12-11 | 2018-12-11 | A kind of image group credit analysis model building method and analysis method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109583447A true CN109583447A (en) | 2019-04-05 |
Family
ID=65928166
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811512310.6A Pending CN109583447A (en) | 2018-12-11 | 2018-12-11 | A kind of image group credit analysis model building method and analysis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109583447A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110265095A (en) * | 2019-05-22 | 2019-09-20 | 首都医科大学附属北京佑安医院 | For HCC recurrence and construction method and the application of the prediction model and nomogram of RFS |
CN111814868A (en) * | 2020-07-03 | 2020-10-23 | 苏州动影信息科技有限公司 | Model based on image omics feature selection, construction method and application |
CN111951962A (en) * | 2020-07-27 | 2020-11-17 | 中山大学孙逸仙纪念医院 | Bevacizumab curative effect prediction model for treating radioactive brain injury and construction method thereof |
CN112262440A (en) * | 2020-07-15 | 2021-01-22 | 北京肿瘤医院(北京大学肿瘤医院) | Method and system for judging cancer treatment response through image omics characteristics |
CN113344853A (en) * | 2021-05-06 | 2021-09-03 | 东南大学 | Robustness detection method based on random displacement algorithm brain image omics |
CN113362932A (en) * | 2021-06-24 | 2021-09-07 | 河北大学 | Image omics feature screening method based on CC-attention mechanism |
CN116230237A (en) * | 2023-05-06 | 2023-06-06 | 四川省医学科学院·四川省人民医院 | Lung cancer influence evaluation method and system based on ROI focus features |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101584578A (en) * | 2008-09-10 | 2009-11-25 | 上海市肺科医院 | Analysis method of comprehensive grading parameters for sarcoidosis and atypical tuberculosis |
CN102512244A (en) * | 2009-06-20 | 2012-06-27 | 上海市肺科医院 | Method for analyzing overall rating parameters of sarcoidosis and atypical tuberculosis |
CN107326090A (en) * | 2017-08-23 | 2017-11-07 | 武汉大学 | Quantitative detecting method for the blood platelet LncRNA of Diagnosis of Non-Small Cell Lung |
CN107368695A (en) * | 2017-08-24 | 2017-11-21 | 南方医科大学南方医院 | The construction method of GISTs malignant potential disaggregated model based on radiation group |
US20180214111A1 (en) * | 2015-08-06 | 2018-08-02 | Case Western Reserve University | Decision support for disease characterization and treatment response with disease and peri-disease radiomics |
US20180276498A1 (en) * | 2017-03-21 | 2018-09-27 | Case Western Reserve University | Predicting prostate cancer recurrence in pre-treatment prostate magnetic resonance imaging (mri) with combined tumor induced organ distension and tumor radiomics |
CN108776962A (en) * | 2018-04-11 | 2018-11-09 | 浙江师范大学 | A method of the structure good pernicious prediction model of lung neoplasm |
CN108897984A (en) * | 2018-05-07 | 2018-11-27 | 上海理工大学 | Based on correlation analysis between CT images group feature and lung cancer gene expression |
CN108898160A (en) * | 2018-06-01 | 2018-11-27 | 中国人民解放军战略支援部队信息工程大学 | Breast cancer tissue's Pathologic Grading method based on CNN and image group Fusion Features |
-
2018
- 2018-12-11 CN CN201811512310.6A patent/CN109583447A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101584578A (en) * | 2008-09-10 | 2009-11-25 | 上海市肺科医院 | Analysis method of comprehensive grading parameters for sarcoidosis and atypical tuberculosis |
CN102512244A (en) * | 2009-06-20 | 2012-06-27 | 上海市肺科医院 | Method for analyzing overall rating parameters of sarcoidosis and atypical tuberculosis |
US20180214111A1 (en) * | 2015-08-06 | 2018-08-02 | Case Western Reserve University | Decision support for disease characterization and treatment response with disease and peri-disease radiomics |
US20180276498A1 (en) * | 2017-03-21 | 2018-09-27 | Case Western Reserve University | Predicting prostate cancer recurrence in pre-treatment prostate magnetic resonance imaging (mri) with combined tumor induced organ distension and tumor radiomics |
CN107326090A (en) * | 2017-08-23 | 2017-11-07 | 武汉大学 | Quantitative detecting method for the blood platelet LncRNA of Diagnosis of Non-Small Cell Lung |
CN107368695A (en) * | 2017-08-24 | 2017-11-21 | 南方医科大学南方医院 | The construction method of GISTs malignant potential disaggregated model based on radiation group |
CN108776962A (en) * | 2018-04-11 | 2018-11-09 | 浙江师范大学 | A method of the structure good pernicious prediction model of lung neoplasm |
CN108897984A (en) * | 2018-05-07 | 2018-11-27 | 上海理工大学 | Based on correlation analysis between CT images group feature and lung cancer gene expression |
CN108898160A (en) * | 2018-06-01 | 2018-11-27 | 中国人民解放军战略支援部队信息工程大学 | Breast cancer tissue's Pathologic Grading method based on CNN and image group Fusion Features |
Non-Patent Citations (5)
Title |
---|
CUI YANFEN ETC.: "Radiomics analysis of multiparametric MRI for prediction of pathological complete response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer", 《EUROPEAN RADIOLOGY》 * |
LIU ZHENYU ETC.: "Radiomics Analysis for Evaluation of Pathological Complete Response to Neoadjuvant Chemoradiotherapy in Locally Advanced Rectal Cancer", 《CLINICAL CANCER RESEARCH》 * |
何兰等: "CT影像组学在非小细胞肺癌临床分期中的价值", 《中华放射学杂志》 * |
温强: "影像组学在预测胸部肿瘤异质性及放疗疗效中的应用", 《万方数据库.学位论文库》 * |
董玲: "基于影像组学的肺癌ALK突变患者预后分析", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110265095A (en) * | 2019-05-22 | 2019-09-20 | 首都医科大学附属北京佑安医院 | For HCC recurrence and construction method and the application of the prediction model and nomogram of RFS |
CN111814868A (en) * | 2020-07-03 | 2020-10-23 | 苏州动影信息科技有限公司 | Model based on image omics feature selection, construction method and application |
CN112262440A (en) * | 2020-07-15 | 2021-01-22 | 北京肿瘤医院(北京大学肿瘤医院) | Method and system for judging cancer treatment response through image omics characteristics |
WO2022011616A1 (en) * | 2020-07-15 | 2022-01-20 | 北京肿瘤医院(北京大学肿瘤医院) | Method and system for determining cancer therapy reaction by means of radiomics feature |
CN111951962A (en) * | 2020-07-27 | 2020-11-17 | 中山大学孙逸仙纪念医院 | Bevacizumab curative effect prediction model for treating radioactive brain injury and construction method thereof |
CN111951962B (en) * | 2020-07-27 | 2022-03-11 | 中山大学孙逸仙纪念医院 | Bevacizumab curative effect prediction model for treating radioactive brain injury and construction method thereof |
CN113344853A (en) * | 2021-05-06 | 2021-09-03 | 东南大学 | Robustness detection method based on random displacement algorithm brain image omics |
CN113344853B (en) * | 2021-05-06 | 2024-03-12 | 东南大学 | Robustness detection method based on random displacement algorithm brain image histology |
CN113362932A (en) * | 2021-06-24 | 2021-09-07 | 河北大学 | Image omics feature screening method based on CC-attention mechanism |
CN116230237A (en) * | 2023-05-06 | 2023-06-06 | 四川省医学科学院·四川省人民医院 | Lung cancer influence evaluation method and system based on ROI focus features |
CN116230237B (en) * | 2023-05-06 | 2023-07-21 | 四川省医学科学院·四川省人民医院 | Lung cancer influence evaluation method and system based on ROI focus features |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109583447A (en) | A kind of image group credit analysis model building method and analysis method | |
US20220323776A1 (en) | Systems and methods for finding regions of interest in hematoxylin and eosin (h&e) stained tissue images and quantifying intratumor cellular spatial heterogeneity in multiplexed/hyperplexed fluorescence tissue images | |
Naqi et al. | Lung nodule detection and classification based on geometric fit in parametric form and deep learning | |
Ahmmed et al. | Classification of tumors and it stages in brain MRI using support vector machine and artificial neural network | |
Hu et al. | Reproducibility with repeat CT in radiomics study for rectal cancer | |
Lee et al. | Random forest based lung nodule classification aided by clustering | |
EP2391887B1 (en) | Image-based risk score-a prognostic predictor of survival and outcome from digital histopathology | |
Albayrak et al. | Automatic cell segmentation in histopathological images via two-staged superpixel-based algorithms | |
Nanda et al. | Computer-aided segmentation of liver lesions in CT scans using cascaded convolutional neural networks and genetically optimised classifier | |
JP2023507109A (en) | Automated tumor identification and segmentation from medical images | |
US20200058125A1 (en) | Comparative cancer survival models to assist physicians to choose optimal treatment | |
CN116188423B (en) | Super-pixel sparse and unmixed detection method based on pathological section hyperspectral image | |
Kohad et al. | Application of machine learning techniques for the diagnosis of lung cancer with ANT colony optimization | |
Kaur et al. | A novel fully automatic multilevel thresholding technique based on optimized intuitionistic fuzzy sets and tsallis entropy for MR brain tumor image segmentation | |
Napoleon et al. | Classifying lung cancer as benign and malignant nodule using ann of back-propagation algorithm and glcm feature extraction on chest x-ray images | |
CN113270191A (en) | Data correction and classification method and storage medium | |
Defeudis et al. | Comparison of radiomics approaches to predict resistance to 1st line chemotherapy in liver metastatic colorectal cancer | |
CN116229176A (en) | Medical image classification method and device combining image histology characteristics and clinical characteristics | |
Çetin et al. | Fuzzy local information c-means algorithm for histopathological image segmentation | |
CN115457069A (en) | Image-based feature extraction and prognosis model establishment method and device | |
Mohamed et al. | A Review of Breast Cancer Classification and Detection Techniques | |
Hibbard | Region segmentation using information divergence measures | |
Panic et al. | A fully automatic deep learning algorithm to segment rectal Cancer on MR images: a multi-center study | |
Brunner et al. | Toward unsupervised classification of calcified arterial lesions | |
Athanasiadis et al. | Segmentation of complementary DNA microarray images by wavelet-based Markov random field model |
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: 20190405 |
|
RJ01 | Rejection of invention patent application after publication |