CN101976303A - Breast mass and calcific benign-malignant automatic recognition and quantitative image evaluation system - Google Patents
Breast mass and calcific benign-malignant automatic recognition and quantitative image evaluation system Download PDFInfo
- Publication number
- CN101976303A CN101976303A CN2010105149211A CN201010514921A CN101976303A CN 101976303 A CN101976303 A CN 101976303A CN 2010105149211 A CN2010105149211 A CN 2010105149211A CN 201010514921 A CN201010514921 A CN 201010514921A CN 101976303 A CN101976303 A CN 101976303A
- Authority
- CN
- China
- Prior art keywords
- calcification
- breast
- lump
- clinical
- parameter
- 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.)
- Granted
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 29
- 206010006272 Breast mass Diseases 0.000 title claims description 8
- 206010028980 Neoplasm Diseases 0.000 claims abstract description 38
- 238000004458 analytical method Methods 0.000 claims abstract description 33
- 210000000481 breast Anatomy 0.000 claims abstract description 27
- ZOKXTWBITQBERF-UHFFFAOYSA-N Molybdenum Chemical compound [Mo] ZOKXTWBITQBERF-UHFFFAOYSA-N 0.000 claims abstract description 24
- 229910052750 molybdenum Inorganic materials 0.000 claims abstract description 24
- 239000011733 molybdenum Substances 0.000 claims abstract description 24
- 208000026310 Breast neoplasm Diseases 0.000 claims abstract description 18
- 238000009792 diffusion process Methods 0.000 claims abstract description 18
- 230000005311 nuclear magnetism Effects 0.000 claims abstract description 10
- 206010006187 Breast cancer Diseases 0.000 claims abstract description 9
- 238000013499 data model Methods 0.000 claims abstract description 9
- 238000000034 method Methods 0.000 claims abstract description 7
- 238000005516 engineering process Methods 0.000 claims abstract description 6
- 210000002445 nipple Anatomy 0.000 claims abstract description 4
- 230000002308 calcification Effects 0.000 claims description 60
- 239000004576 sand Substances 0.000 claims description 20
- 210000005075 mammary gland Anatomy 0.000 claims description 18
- 230000000505 pernicious effect Effects 0.000 claims description 16
- 230000001575 pathological effect Effects 0.000 claims description 11
- 238000011002 quantification Methods 0.000 claims description 11
- 230000012010 growth Effects 0.000 claims description 9
- 230000007170 pathology Effects 0.000 claims description 9
- 230000004614 tumor growth Effects 0.000 claims description 9
- 206010006223 Breast discharge Diseases 0.000 claims description 8
- 238000012937 correction Methods 0.000 claims description 6
- 238000001574 biopsy Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000000877 morphologic effect Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000003705 background correction Methods 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 238000013178 mathematical model Methods 0.000 claims description 2
- 230000003902 lesion Effects 0.000 abstract description 9
- 238000003745 diagnosis Methods 0.000 abstract description 5
- 238000010191 image analysis Methods 0.000 abstract description 2
- 238000012216 screening Methods 0.000 abstract description 2
- 238000005056 compaction Methods 0.000 abstract 1
- 208000004434 Calcinosis Diseases 0.000 description 44
- 238000003759 clinical diagnosis Methods 0.000 description 13
- 230000004069 differentiation Effects 0.000 description 9
- 201000011510 cancer Diseases 0.000 description 5
- 238000005481 NMR spectroscopy Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 230000003211 malignant effect Effects 0.000 description 3
- 238000004445 quantitative analysis Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000002059 diagnostic imaging Methods 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 208000004396 mastitis Diseases 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 206010048782 Breast calcifications Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 208000030270 breast disease Diseases 0.000 description 1
- 201000010983 breast ductal carcinoma Diseases 0.000 description 1
- 230000010261 cell growth Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000004195 computer-aided diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004992 fission Effects 0.000 description 1
- 230000007773 growth pattern Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000008267 milk Substances 0.000 description 1
- 210000004080 milk Anatomy 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
Images
Landscapes
- Magnetic Resonance Imaging Apparatus (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention provides a breast lesion quantitative image evaluation (CAD) system and an application method thereof. The breast lesion quantitative image evaluation system adopts image data of a fractal technology, pattern analysis, and the like to extract an excavating mean and a mathematical modeling algorithm, combines clinical data to establish a breast lesion grown diffusion nonlinear data model and is applied to tumor medical image analysis and tumor disease risk evaluation. The nonlinear data model comprises the clinical parameters of breast tumor grown diffusion state characteristic parameters, calcific state characteristic parameters, breast surface characteristic limitation asymmetric compaction comparison, nipple retraction, pachyderma, structural distortion, and the like. The breast lesion quantitative image evaluation system has a full-graphical interface, can lead a molybdenum target, nuclear magnetism and ultrasonic image data in and is convenient and quick in operation (one-key operation). By the CAD system, a benign-malignant forecasting numerical value and a tumor classification forecasting value of breast molybdenum target piece (nuclear magnetism and ultrasonic) lesions can be calculated, and results of the benign-malignant forecasting numerical value and the tumor classification forecasting value can be applied to breast image auxiliary diagnosis and breast filming screening.
Description
Technical field
The invention belongs to mastotic medical diagnostic equipment, particularly relate to a kind of breast lump, calcification and very dislike identification and quantification image (molybdenum target, ultrasonic, nuclear-magnetism) evaluation system automatically, comprise the image data segmented extraction and manage modeling based on the number of clinical pathology image data data; System of the present invention is applicable to general clinical assistant diagnosis and generaI investigation.
Background technology
Breast image learn to check, particularly breast molybdenum target take the photograph sheet for the diagnosis of tumor of breast, by stages, the judgement of therapeutic evaluation provides important evidence, has been applied to breast screening.Regrettably in current clinical diagnosis, conventional diagnostic imaging only is confined to measurement and some qualitative evaluations of the tumour scale size and the simple shape factor, even some can use external image processing software Image-Pro Plus to carry out the doctor or the medical matters scientific research personnel of manual calculations, because separately to the processing horizontal difference of picture, even same people's different operating number of times, the result difference that obtains is bigger.
Along with improving constantly of medical image quantitative analysis tech, in the fundamental research of tumour, also need to estimate the growth and the diffusion characteristic of different tumour classifications by image quantitative analysis.Particularly fractal notion is accepted by more researcher, as how much FRACTAL DIMENSION of tumor boundaries to estimate tumour and in grown-diffused process, exchange with the nutrition of its perienchyma and degree of opening very important; Complicated FRACTAL DIMENSION then can disclose the important information of swollen thing growth inside fission such as heterogeneity etc.This The Application of Technology has demonstrated its original value.Fractal technology and relevant pattern analysis means are applied to the level of significance assessment of the analysis of kinds of tumors medical image and the tumour state of an illness gradually, and have obtained some and tumour fundamental research and clinical diagnosis are had the result of important references value.From the angle of fractal mathematics and nonlinear physics, the division of inside tumor cancer cell filial generation, developing can be by complicated fractal characterization, and the fractal dimension of tumor boundaries profile is the complicacy that characterizes the boundary profile after the tumor growth diffusion.
Innocent and malignant tumour exists different diffusions and growth pattern.The present invention examines or check tumor of breast growth diffusion parameter, complicated FRACTAL DIMENSION, heterogeneity and enclosed mass degree etc. as boundary profile FRACTAL DIMENSION, inside tumor, and comprehensive breast lesion calcification feature, and Clinical symptoms sex character resembles, a kind of succinct effectively practicality tumor imaging quantitative estimation method is provided, calculate the good pernicious predicted numerical value of tumour, for clinical reference.
Summary of the invention
The objective of the invention is to overcome the problem that existing breast molybdenum target sheet (nuclear-magnetism, ultrasonic) diagnostic imaging lacks the quantitative evaluation index, a kind of mammary gland affection quantification image auxiliary diagnosis evaluation system is provided, this quantification image evaluation system can carry out quantitative analysis to breast molybdenum target sheet (nuclear-magnetism, ultrasonic) image, provides breast lesion good pernicious predicted value.Computer-aided diagnosis of the present invention (CAD) system adopts operating in a key to the identification of lump, calcification, and is convenient and swift, to the operational computations result of same picture, can not vary with each individual.
The technical solution used in the present invention is as follows:
A kind of mammary gland affection quantification image evaluation system, it adopts fractal technology and pattern analysis means to be applied in the level of significance assessment of the analysis of tumour medical image and the tumour state of an illness, set up and adopted the nonlinear data model of breast lesion cell growth diffusion, described nonlinear data model has comprised tumor of breast growth diffusion form characteristic parameter, calcification pattern characteristic parameter and clinical parameter.
In the technique scheme, the weight ratio of tumor growth diffusion form characteristic parameter is 0.1～1.0, and the weight ratio of calcification pattern characteristic parameter is 0.001～1.0; The weight ratio of clinical pathology value parameter is 0.1～1.0.
Described tumor growth diffusion form characteristic parameter mainly comprises the complicated fractal M of tumour
F, tumor boundaries how much FRACTAL DIMENSION D
FClinical parameter comprises patient's age Y, nipple discharge U, mammary gland medical history and nurture history V.
Described calcification pattern characteristic parameter comprises that mainly the calcification density P of calcification spot, the minimum boundary rectangle frame of calcification spot estimate S, grains of sand calcified plaque ratio N
S
Described clinical parameter mainly comprises the clinical diagnosis classification Y of patient age y, the state of an illness
C
Nonlinear data model of the present invention comprises that multifactor linear fit returns mathematical modeling, by tumor growth diffusion form characteristic parameter is carried out different weight analysis with clinical parameter, filter out the swollen thing classification that can meet clinical state of an illness data and judge regression equation (1):
Y
E＝a*D
F+b*M
F+c*y+d*P+e*H (1)
Wherein a, b, c, d, e are for returning weight coefficient, Y
EIt is the predictability pathological grading.
In conjunction with clinical parameter to M
F, D
FDo following correction, obtain following fit equation (2) and (3):
M
mul＝y*M
F+f，(2)
F
simp＝y*D
F+g，(3)
Wherein, M
MulBe through revised complicated FRACTAL DIMENSION; F
SimpBe through revised border FRACTAL DIMENSION; Y is the age correction factor, and f, g are the parameters that obtains with a large amount of underwent operative biopsy image data regretional analysis that this CAD calculates, and y is the patient age parameter.
Obtain biopsy obtains based on underwent operative the good evil threshold data and the fit equation (4) and the logical equatiion (5) of clinical image morphological data:
P＝N
ghb/S；(4)
E＝M
mul|F
simp|y*P&Ns；(5)
Wherein, N
SBe grains of sand calcified plaque ratio; P is the calcification density of calcified plaque; E is that predictability pathology is very disliked.
CAD system concrete operations of the present invention comprise the steps:
(1) clinical breast molybdenum target (nuclear-magnetism, ultrasonic) is taken the photograph sheet and carry out pre-service, unified image is ash 8 degree, unified background correction and background compensation, the false calcification of unified filtering skin edge, false lump, automated manner obtains focal zone/region of interest that breast molybdenum target (nuclear-magnetism, ultrasonic) is taken the photograph sheet;
(2) tumor growth that calculates through pretreated region of interest spreads the form characteristic parameter, and swollen thing parameter comprises the geometry FRACTAL DIMENSION D on the border between swollen thing and the normal structure
F, lump inside complicated fractal M
F, swollen thing inside heterogeneous H, lump equivalent diameter D; The calcification characteristic parameter comprises that mainly the calcification density P of calcification spot, the minimum boundary rectangle frame of calcification spot estimate S, grains of sand calcified plaque ratio N
SAnd mammary gland exterior appearance image parameters such as the asymmetric fine and close comparison of the limitation of the feature of mammary gland exterior appearance, nipple retraction, pachyderma, structural distortion;
(3) introduce age y, nipple discharge U, the mammary gland case history V clinical parameter that comprises patient, and in conjunction with the focus image feature parameter and the mammary gland exterior appearance image parameters that extract, calculate based on the mathematical model of being built, provide the predictability state of an illness and analyze and good pernicious judgement by stages.
CAD system of the present invention has following beneficial effect:
Mammary gland affection quantification image evaluation system of the present invention is the full graphics interface; Can import molybdenum target, nuclear-magnetism, ultrasonic image data and easy to operate (key operation); The non-linear data model of breast lesion growth diffusion has comprised tumor of breast growth diffusion quantitative parameter, calcification parameter and clinical parameter.In the modeling of number reason, the present invention has adopted the AI programing technique with learning functionality, can further improve the judgement precision of CAD system of the present invention by regularly newly-increased case being screened in conjunction with clinical effectiveness.
The present invention calculates the good pernicious predicted numerical value of breast molybdenum target sheet (comprising nuclear magnetic resonance and ultrasonic image) pathology, not only can be applicable to mammary gland disease image clinical assistant diagnosis, also can be used for taking the photograph the sheet generaI investigation based on the mammary gland of molybdenum target image.
Description of drawings
Figure 1A, Figure 1B and Fig. 1 C are respectively the molybdenum target sheet of embodiment and discern the lump exemplary plot automatically.
Fig. 2 A and Fig. 2 B are respectively simple property calcification molybdenum target figure.
Fig. 3 A and Fig. 3 B are respectively ultrasonic identification lump embodiment exemplary plot.
Fig. 4 A and Fig. 4 B are respectively nuclear magnetic resonance MRI sheet identification lump exemplary plot.
Embodiment
The present invention is described further below by embodiment.
Wherein embodiment 1～9, relates to the application content of operation concrete steps (1)～(3) scheme, quotes each figure and does complementary explanation.Embodiment 1 adds the graphical analysis of calcified plaque compound type molybdenum target for lump; Embodiment 2 is the graphical analysis of simple property calcified plaque molybdenum target; Embodiment 3 and 4 is the molybdenum target graphical analysis; Embodiment 6 and 7 is the ultrasonoscopy analysis; Embodiment 8 and 9 is the nuclear magnetic resonance image analysis.
Embodiment 1
Patient A; Age: 80 years old; Clinical diagnosis: 2 grades of no nipple discharge malignant breast carcinomas
Use CAD system, obtain breast molybdenum target automatically and take the photograph the optimum of sheet and malignant tumour region of interest.Figure 1A, Figure 1B and Fig. 1 C are respectively that the molybdenum target sheet is discerned the lump exemplary plot automatically, left figure is original molybdenum target sheet among the figure, wherein the red curve area surrounded is the pernicious lump that CAD automatically identifies, the middle graph that is stitched together with left figure is the calcified regions sectional drawing that amplifies, right figure is the calcified plaque that CAD extracts automatically, and yellow rectangle frame is the boundary rectangle frame of calcified regions.When no yellow rectangle frame occurs automatically, showing does not have calcified plaque (as Fig. 1 C) in the image, in case yellow rectangle frame occurs automatically, show the existence that the Microcalcification spot is arranged, even do not observe tangible calcified plaque (as Figure 1B) in the yellow rectangle frame, illustrate that near the superfine little grains of sand calcified plaque that has occurred the pixel dimension caught by this CAD system, but too tiny and can not be discovered by naked eyes.The whole image parameters values extracted of excavating are positioned at CAD system graphical interfaces bottom section, and the evaluation reference that draws according to interior established model is positioned at zone, the lower right corner, CAD system graphical interfaces bottom.
Calculate lump parameter among Figure 1A: FRACTAL DIMENSION D according to CAD
F=1.2033, heterogeneous H=0.3762, multifractal dimension M
F=0.0909, calcification parameter: calcification density P=0.0914, calcified plaque number N=7, grains of sand calcified plaque Ns=9.324%.
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Figure 1A is judged as 2 grades of pernicious right breast cancer by CAD, and the image evaluation result conforms to pathological examination.
Embodiment 2
Patient B; Age: 75 years old; Clinical diagnosis: 3 grades of nipple discharge malignant breast carcinomas are arranged
Use CAD system, obtain breast molybdenum target automatically and take the photograph the optimum of sheet and malignant tumour region of interest.Figure 1A described in the handled figure similar embodiment 1.Left figure is former molybdenum target sheet, and the centre is the CAD enlarged drawing of the pernicious calcified plaque of identification automatically, and yellow rectangle frame is the minimum boundary rectangle frame of calcified plaque.The whole image parameters values extracted of excavating are positioned at CAD system graphical interfaces bottom section, and the evaluation reference that draws according to interior established model is positioned at zone, the lower right corner, CAD system graphical interfaces bottom.
Calculate lump parameter among Figure 1B: FRACTAL DIMENSION D according to CAD
F=1.2459, heterogeneous H=0.4735, multifractal dimension M
F=0.1364, calcification parameter: calcification density P=0.0378, calcified plaque number N=14, grains of sand calcified plaque Ns=12.066%
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Figure 1B is judged as 3 grades of pernicious right breast cancer by CAD, and the image evaluation result conforms to pathological examination.
Embodiment 3
Patient C; Age: 44 years old; Clinical diagnosis: touch painful mastitis
Calculate lump parameter among Fig. 1 C: FRACTAL DIMENSION D according to CAD
F=0, heterogeneous H=0, multifractal dimension M
F=0, the calcification parameter: calcification density P=0, calcified plaque number N=0, grains of sand calcified plaque Ns=0,
Because the CAD check result is not have lump and calcified plaque, so result of calculation is 0 entirely, judged result is not for finding pernicious lump and calcification, and (mastitis is common general mastosis) conforms to clinical diagnosis
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Fig. 1 C is judged as by CAD and is not found pernicious lump and calcification, and the image evaluation result conforms to pathological examination.
Embodiment 4
Patient D; Age: 33 years old; Clinical diagnosis: right newborn nipple discharge, duct carcinoma
Calculate lump parameter among Fig. 2 A: FRACTAL DIMENSION D according to CAD
F=0, heterogeneous H=0, multifractal dimension M
F=0, the calcification parameter: calcification density P=0.0901, calcified plaque number N=30, grains of sand calcified plaque Ns=8.894%,
Because the CAD check result is not have lump, so lump result of calculation is 0 entirely, only judges with the calcification data.
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Fig. 2 A is judged as the newborn breast cancer in the pernicious right side by CAD, and the image evaluation result conforms to pathological examination.
Patient D; Age: 33 years old; Clinical diagnosis: left side breast is normal
Calculate lump parameter among Fig. 2 B: FRACTAL DIMENSION D according to CAD
F=0, heterogeneous H=0, multifractal dimension M
F=0, the calcification parameter: calcification density P=0, calcified plaque number N=0, grains of sand calcified plaque Ns=0,
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Automatic analysis and judgment based on above-mentioned CAD system, Fig. 2 B is judged as left breast by CAD and is not found pernicious lump and calcification, and this example and embodiment 4 are same patient, and this patients clinical diagnostic result is that right milk duct cancer is pernicious, left side breast is normal, and the image evaluation result conforms to pathological examination.
Embodiment 6
Patient E; Age: 46 years old; Clinical diagnosis: breast cancer, 2 grades
Automatically identify lump according to CAD and calculate lump parameter among Fig. 3 A: FRACTAL DIMENSION D
F=1.1844, heterogeneous H=0.8792, multifractal dimension M
F=0.7812, calcification parameter: calcification density P=0, calcified plaque number N=0, grains of sand calcified plaque Ns=0.Owing to ultrasonicly Microcalcification is detected insensitive, monitors so ultrasonic picture is only done lump.
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logic journey (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Fig. 3 A is judged as 2 grades of malignant breast carcinomas by CAD, and the image evaluation result conforms to pathological examination.
Embodiment 7
Patient F; Age: 53; Clinical diagnosis: benign tumour
Automatically identify lump according to CAD and calculate lump parameter among Fig. 3 B: FRACTAL DIMENSION D
F=1.1123, heterogeneous H=0.7690, multifractal dimension M
F=0.9711, calcification parameter: calcification density P=0, calcified plaque number N=0, grains of sand calcified plaque Ns=0.Owing to ultrasonicly Microcalcification is detected insensitive, monitors so ultrasonic picture is only done lump.
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logic journey (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Fig. 3 B is judged as benign tumour by CAD, and the image evaluation result conforms to pathological examination.
Embodiment 8
Patient G; Age: 66; Clinical diagnosis: breast cancer, pernicious 1 grade
Automatically identify lump according to CAD and calculate lump parameter among Fig. 4 A: FRACTAL DIMENSION D
F=1.1759, heterogeneous H=0.8235, multifractal dimension M
F=0.3496, the calcification parameter: calcification density P=0, calcified plaque number N=0, grains of sand calcified plaque Ns=0 because that MRI detects Microcalcification is insensitive, monitors so the MRI picture is only done lump.
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Fig. 4 A is judged as the newborn breast cancer in the pernicious right side by CAD, and the image evaluation result conforms to pathological examination.
Embodiment 9
Patient H; Age: 52 years old; Clinical diagnosis: optimum
Automatically identify lump according to CAD and calculate lump parameter among Fig. 4 B: FRACTAL DIMENSION D
F=1.1568, heterogeneous H=0.8276, multifractal dimension M
F=0.3612, the calcification parameter: calcification density P=0, calcified plaque number N=0, grains of sand calcified plaque Ns=0 because that MRI detects Microcalcification is insensitive, monitors so the MRI picture is only done lump.
Judge regression equation (1): Y according to classification
E=a*D
F+ b*M
F+ c*y+d*P+e*H very dislikes differentiation logical equatiion (5): E=M according to tumour
Mul| F
Simp| y*P﹠amp; Ns, Conjoint Analysis is differentiated.Based on the automatic analysis and judgment of above-mentioned CAD system, Fig. 4 B is judged as benign mastopathy disease by CAD and becomes, and the image evaluation result conforms to pathological examination.
The above, it only is preferred embodiment of the present invention, be not that the present invention is done any pro forma restriction, so every technical solution of the present invention content that do not break away from,, all still belong in the scope of technical scheme that claims of the present invention limit any simple modification, equivalent variations and modification that above embodiment did according to technical spirit of the present invention.
Claims (5)
1. a breast lump, the good evil of calcification are discerned and the quantification image evaluation system automatically, it adopts fractal technology and pattern analysis means to be applied in the level of significance assessment of the analysis of tumour medical image and the tumour state of an illness, sets up and adopted the nonlinear data model of tumor of breast pathology growth diffusion.Described nonlinear data model has comprised the complicated FRACTAL DIMENSION M of tumor of breast growth diffusion morphological feature
F, lump boundary geometrical FRACTAL DIMENSION D
F, lump inner heterogeneous H, calcification pattern characteristic parameter P, and clinical parameter y, nipple discharge U, mammary gland medical history and nurture history V.It is characterized in that:
The weight ratio of described tumor growth diffusion form characteristic parameter is 0.1～1.0; The weight ratio of calcification pattern characteristic parameter is 0.001～1.0; The weight ratio of clinical pathology value parameter is 0.1～1.0.Described tumor growth diffusion form characteristic parameter comprises the inner complicated fractal M of lump
F, lump boundary geometrical FRACTAL DIMENSION D
F, the inner heterogeneous H of lump; Described tumour calcification pattern characteristic parameter P; Described clinical parameter comprises patient's age y, nipple discharge U, mammary gland medical history and nurture history V.Described nonlinear data model comprises that multifactor linear fit returns mathematical modeling, by tumor growth diffusion form characteristic parameter is carried out different weight analysis with clinical parameter, filter out the regression equation that can meet clinical state of an illness data, obtain based on clinical pathology classification Y
CJudge regression equation with the breast tumor of clinical image morphological data:
Y
E＝a*D
F+b*M
F+c*y+d*P+e*H；(1)
Wherein a, b, c, d, e are for returning weight coefficient, Y
EIt is the predictability pathological grading.To M
F, D
FDo following correction,
M
mul＝y*M
F+f (2)
F
simp＝y*D
F+g，?(3)
Wherein, M
MulBe through revised complicated FRACTAL DIMENSION; F
SimpBe through revised border FRACTAL DIMENSION; Y is the age correction factor, and f, g select the statistical study corresponding data that calculates through the mammary gland picture that molybdenum target, MRI, the corresponding underwent operative of ultrasonic device or biopsy confirm.
2. breast lump according to claim 1, good identification and the quantification image evaluation system automatically of disliking of calcification, it is characterized in that: described lump characteristic parameter has the complicated FRACTAL DIMENSION M of the swollen thing growth inside heterogeneity of reflection
F, the swollen coarse degree of opening of the thing boundary profile border FRACTAL DIMENSION D of reflection
F, reflection calcification pattern characteristic density distribution parameter P, grains of sand calcified plaque ratio Ns and mammary gland exterior appearance the clinical parameters such as the asymmetric fine and close comparison of limitation, nipple retraction, pachyderma, structural distortion of feature; Described calcification pattern characteristic parameter comprises: the calcification density P of calcified plaque, calcified plaque number N
GhbWith grains of sand calcified plaque ratio N
SCalcified plaque minimum diameter and calcification mean diameter estimate D
Min, D
Mean, the minimum boundary rectangle frame of the complete area that calcified plaque distributes is estimated S; Judge fit equation by the calcified plaque that the calcification pattern characteristic parameter is carried out different weight analysis with clinical parameter, filter out meeting clinical state of an illness data:
P＝N
ghb/S (4)。
3. breast lump according to claim 2, good identification and the quantification image evaluation system automatically of disliking of calcification, it is characterized in that: described lump, calcified plaque judge that fit equation is to judge the good evil of tumour by following logical expression:
E＝M
mul|F
simp|y*P&N
S，(5)
Wherein, y is the age correction factor, M
MulBe through revised complicated FRACTAL DIMENSION; F
SimpBe through revised border FRACTAL DIMENSION; N
SIt is grains of sand calcified plaque ratio; P is the calcification density of calcified plaque; E is that predictability pathology is very disliked.
4. the application process of a breast lump, the automatic identification of the good evil of calcification and quantification image evaluation system is characterized in that comprising the steps:
(1) clinical breast molybdenum target (nuclear-magnetism, ultrasonic) is taken the photograph sheet and carry out pre-service, unified image is ash 8 degree, unified background correction and background compensation, the false calcification of unified filtering skin edge, false lump, automated manner obtains focal zone/region of interest that breast molybdenum target (nuclear-magnetism, ultrasonic) is taken the photograph sheet;
(2) tumor growth that calculates through pretreated region of interest spreads the form characteristic parameter, and swollen thing parameter comprises the geometry FRACTAL DIMENSION D on the border between swollen thing and the normal structure
F, lump inside complicated fractal M
F, swollen thing inside heterogeneous H, lump equivalent diameter D; The calcification characteristic parameter comprises that mainly the calcification density P of calcification spot, the minimum boundary rectangle frame of calcification spot estimate S, grains of sand calcified plaque ratio N
SAnd mammary gland exterior appearance image parameters such as the asymmetric fine and close comparison of the limitation of the feature of mammary gland exterior appearance, nipple retraction, pachyderma, structural distortion;
(3) introduce age y, nipple discharge U, the mammary gland case history V clinical parameter that comprises patient, and in conjunction with the focus image feature parameter and the mammary gland exterior appearance image parameters that extract, calculate based on the mathematical model of being built, provide the predictability state of an illness and analyze and good pernicious judgement by stages.
5. breast lump according to claim 4, the good application process of disliking identification automatically and quantification image evaluation system of calcification is characterized in that: judge the good logical equatiion of disliking of breast cancer:
E＝M
mul|F
simp|y*P&N
S
Wherein, y is the age correction factor, M
MmlBe through revised complicated FRACTAL DIMENSION; F
SimpBe through revised border FRACTAL DIMENSION; N
SIt is grains of sand calcified plaque ratio; P is the calcification density of calcified plaque; E is that predictability pathology is very disliked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010514921 CN101976303B (en) | 2010-10-21 | 2010-10-21 | Breast mass and calcific benign-malignant automatic recognition and quantitative image evaluation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010514921 CN101976303B (en) | 2010-10-21 | 2010-10-21 | Breast mass and calcific benign-malignant automatic recognition and quantitative image evaluation system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101976303A true CN101976303A (en) | 2011-02-16 |
| CN101976303B CN101976303B (en) | 2013-09-04 |
Family
ID=43576188
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201010514921 Expired - Fee Related CN101976303B (en) | 2010-10-21 | 2010-10-21 | Breast mass and calcific benign-malignant automatic recognition and quantitative image evaluation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101976303B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789536A (en) * | 2011-05-20 | 2012-11-21 | 中国人民解放军第二军医大学 | Method for establishing noninvasive evaluation model for liver surgical treatment risks |
| CN106466188A (en) * | 2015-08-20 | 2017-03-01 | 通用电气公司 | For the quantitative system and method for emission tomography imaging |
| CN107977956A (en) * | 2017-11-17 | 2018-05-01 | 深圳蓝韵医学影像有限公司 | Detect the method, apparatus and computer-readable storage medium of X-ray image tissue regions |
| CN108577844A (en) * | 2018-05-18 | 2018-09-28 | 北京先通康桥医药科技有限公司 | The method and system of opening relationships model based on pressure distribution data, storage medium |
| CN110403576A (en) * | 2019-08-01 | 2019-11-05 | 中国医学科学院北京协和医院 | Application of the three-dimensional photoacoustic imaging in tumor of breast points-scoring system |
| CN111275719A (en) * | 2020-01-19 | 2020-06-12 | 北京推想科技有限公司 | Calcification false positive recognition method, device, terminal and medium and model training method and device |
| CN114494263A (en) * | 2022-04-19 | 2022-05-13 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Medical image lesion detection method, system and equipment integrating clinical information |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040258310A1 (en) * | 2003-02-14 | 2004-12-23 | The University Of Chicago | Method and system for fractal-based analysis of medical image texture |
| CN200984178Y (en) * | 2006-08-31 | 2007-12-05 | 深圳市国基科技有限公司 | Digital multifunctional mammary gland imaging system |
| CN101234026A (en) * | 2008-03-07 | 2008-08-06 | 李立 | Mammary gland affection quantification image evaluation system and using method thereof |
-
2010
- 2010-10-21 CN CN 201010514921 patent/CN101976303B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040258310A1 (en) * | 2003-02-14 | 2004-12-23 | The University Of Chicago | Method and system for fractal-based analysis of medical image texture |
| CN200984178Y (en) * | 2006-08-31 | 2007-12-05 | 深圳市国基科技有限公司 | Digital multifunctional mammary gland imaging system |
| CN101234026A (en) * | 2008-03-07 | 2008-08-06 | 李立 | Mammary gland affection quantification image evaluation system and using method thereof |
Non-Patent Citations (2)
| Title |
|---|
| 李婵婵 等: "分形维参数和异质性参数在乳腺良恶性肿瘤钼靶图像分析鉴别诊断中的应用", 《肿瘤学杂志》 * |
| 李文英 等: "乳腺肿瘤血管影像评价方法", 《中国医学影像技术》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789536B (en) * | 2011-05-20 | 2015-11-18 | 中国人民解放军第二军医大学 | The method for building up of operation on liver Operative risk noninvasively estimating model |
| CN102789536A (en) * | 2011-05-20 | 2012-11-21 | 中国人民解放军第二军医大学 | Method for establishing noninvasive evaluation model for liver surgical treatment risks |
| CN106466188B (en) * | 2015-08-20 | 2020-03-17 | 通用电气公司 | System and method for emission tomography quantification |
| CN106466188A (en) * | 2015-08-20 | 2017-03-01 | 通用电气公司 | For the quantitative system and method for emission tomography imaging |
| CN107977956A (en) * | 2017-11-17 | 2018-05-01 | 深圳蓝韵医学影像有限公司 | Detect the method, apparatus and computer-readable storage medium of X-ray image tissue regions |
| CN107977956B (en) * | 2017-11-17 | 2021-10-29 | 深圳蓝韵医学影像有限公司 | Method, apparatus and computer storage medium for detecting tissue region of X-ray image |
| CN108577844A (en) * | 2018-05-18 | 2018-09-28 | 北京先通康桥医药科技有限公司 | The method and system of opening relationships model based on pressure distribution data, storage medium |
| CN110403576A (en) * | 2019-08-01 | 2019-11-05 | 中国医学科学院北京协和医院 | Application of the three-dimensional photoacoustic imaging in tumor of breast points-scoring system |
| CN110403576B (en) * | 2019-08-01 | 2024-01-30 | 中国医学科学院北京协和医院 | Application of three-dimensional photoacoustic imaging in breast tumor scoring system |
| CN111275719A (en) * | 2020-01-19 | 2020-06-12 | 北京推想科技有限公司 | Calcification false positive recognition method, device, terminal and medium and model training method and device |
| CN111275719B (en) * | 2020-01-19 | 2023-04-07 | 推想医疗科技股份有限公司 | Calcification false positive recognition method, device, terminal and medium and model training method and device |
| CN114494263A (en) * | 2022-04-19 | 2022-05-13 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Medical image lesion detection method, system and equipment integrating clinical information |
| CN114494263B (en) * | 2022-04-19 | 2022-06-17 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Medical image lesion detection method, system and equipment integrating clinical information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101976303B (en) | 2013-09-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101976303B (en) | Breast mass and calcific benign-malignant automatic recognition and quantitative image evaluation system | |
| Nagi et al. | Automated breast profile segmentation for ROI detection using digital mammograms | |
| Nayak et al. | Computer-aided diagnosis of cirrhosis and hepatocellular carcinoma using multi-phase abdomen CT | |
| Panigrahi et al. | Ultrasound image segmentation using a novel multi-scale Gaussian kernel fuzzy clustering and multi-scale vector field convolution | |
| Wulandari et al. | Brain tumor segmentation to calculate percentage tumor using MRI | |
| Anter et al. | Automatic computer aided segmentation for liver and hepatic lesions using hybrid segmentations techniques | |
| Tan et al. | Computer-assisted diagnosis of tuberculosis: a first order statistical approach to chest radiograph | |
| Abhisheka et al. | A comprehensive review on breast cancer detection, classification and segmentation using deep learning | |
| CN104268873B (en) | Breast tumor partition method based on nuclear magnetic resonance images | |
| CN101234026B (en) | Mammary gland affection quantification image evaluation system | |
| Pulagam et al. | Automated lung segmentation from HRCT scans with diffuse parenchymal lung diseases | |
| Duarte et al. | Evaluating geodesic active contours in microcalcifications segmentation on mammograms | |
| Lewis et al. | Detection of breast tumor candidates using marker-controlled watershed segmentation and morphological analysis | |
| Memon et al. | Segmentation of lungs from CT scan images for early diagnosis of lung cancer | |
| Deepa et al. | Efficient ROI segmentation of digital mammogram images using Otsu’s N thresholding method | |
| Sahu et al. | A new hybrid approach using fuzzy clustering and morphological operations for lung segmentation in thoracic CT images | |
| Anuradha et al. | Detection of oral tumor based on marker-controlled watershed algorithm | |
| He et al. | Breast image pre-processing for mammographic tissue segmentation | |
| Rajinikanth et al. | Thermogram assisted detection and analysis of ductal carcinoma in situ (DCIS) | |
| Bandyopadhyay | Survey on segmentation methods for locating masses in a mammogram image | |
| Kaur et al. | Computer-aided diagnosis of renal lesions in CT images: a comprehensive survey and future prospects | |
| Nayak et al. | Automatic segmentation and breast density estimation for cancer detection using an efficient watershed algorithm | |
| Wang et al. | Mammography visual enhancement in CAD-based breast cancer diagnosis | |
| Mastouri et al. | A morphological operation-based approach for Sub-pleural lung nodule detection from CT images | |
| Yap et al. | Fully automatic lesion boundary detection in ultrasound breast images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130904 Termination date: 20151021 |
|
| EXPY | Termination of patent right or utility model |