WO2010032472A1 - 代替正常脳データベース生成装置 - Google Patents
代替正常脳データベース生成装置 Download PDFInfo
- Publication number
- WO2010032472A1 WO2010032472A1 PCT/JP2009/004693 JP2009004693W WO2010032472A1 WO 2010032472 A1 WO2010032472 A1 WO 2010032472A1 JP 2009004693 W JP2009004693 W JP 2009004693W WO 2010032472 A1 WO2010032472 A1 WO 2010032472A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- brain
- blood flow
- value
- image
- state
- Prior art date
Links
- 210000004556 brain Anatomy 0.000 title claims abstract description 284
- 230000003925 brain function Effects 0.000 claims abstract description 59
- 208000014644 Brain disease Diseases 0.000 claims abstract description 55
- 238000000034 method Methods 0.000 claims abstract description 39
- 230000017531 blood circulation Effects 0.000 claims description 120
- 230000003727 cerebral blood flow Effects 0.000 claims description 76
- 230000007717 exclusion Effects 0.000 claims description 13
- 230000001131 transforming effect Effects 0.000 claims 2
- 201000010099 disease Diseases 0.000 abstract description 18
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 abstract description 18
- HGBLNBBNRORJKI-WCABBAIRSA-N cyclacillin Chemical compound N([C@H]1[C@H]2SC([C@@H](N2C1=O)C(O)=O)(C)C)C(=O)C1(N)CCCCC1 HGBLNBBNRORJKI-WCABBAIRSA-N 0.000 description 22
- 238000012360 testing method Methods 0.000 description 20
- 238000002603 single-photon emission computed tomography Methods 0.000 description 16
- 238000010606 normalization Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 210000000877 corpus callosum Anatomy 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 9
- 238000012545 processing Methods 0.000 description 9
- 101100083446 Danio rerio plekhh1 gene Proteins 0.000 description 7
- 210000001652 frontal lobe Anatomy 0.000 description 7
- 210000003016 hypothalamus Anatomy 0.000 description 7
- 210000000869 occipital lobe Anatomy 0.000 description 7
- 238000002600 positron emission tomography Methods 0.000 description 6
- 210000001103 thalamus Anatomy 0.000 description 6
- 230000009466 transformation Effects 0.000 description 5
- 210000004885 white matter Anatomy 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 230000004060 metabolic process Effects 0.000 description 4
- 208000024827 Alzheimer disease Diseases 0.000 description 3
- 101100129500 Caenorhabditis elegans max-2 gene Proteins 0.000 description 3
- 230000000052 comparative effect Effects 0.000 description 3
- 230000003247 decreasing effect Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 230000004153 glucose metabolism Effects 0.000 description 3
- ZCXUVYAZINUVJD-AHXZWLDOSA-N 2-deoxy-2-((18)F)fluoro-alpha-D-glucose Chemical compound OC[C@H]1O[C@H](O)[C@H]([18F])[C@@H](O)[C@@H]1O ZCXUVYAZINUVJD-AHXZWLDOSA-N 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000008344 brain blood flow Effects 0.000 description 2
- 210000001638 cerebellum Anatomy 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000000692 Student's t-test Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000004884 grey matter Anatomy 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012353 t test Methods 0.000 description 1
- 238000003325 tomography Methods 0.000 description 1
- 230000002861 ventricular Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/50—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications
- A61B6/501—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications for diagnosis of the head, e.g. neuroimaging or craniography
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/02—Arrangements for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
- A61B6/03—Computed tomography [CT]
- A61B6/037—Emission tomography
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/50—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications
- A61B6/507—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications for determination of haemodynamic parameters, e.g. perfusion CT
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
-
- 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
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/772—Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H70/00—ICT specially adapted for the handling or processing of medical references
- G16H70/60—ICT specially adapted for the handling or processing of medical references relating to pathologies
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10108—Single photon emission computed tomography [SPECT]
-
- 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
- G06T2207/30016—Brain
Definitions
- the present invention relates to generation of an alternative normal brain database used for diagnosis of diseases based on brain function images such as PET images and SPECT images.
- positron emission tomography drugs labeled with positron nuclides such as 18 F-FDG are intravenously administered to a subject, and ⁇ -rays as annihilation radiation in each part of the brain are measured to measure sugar.
- positron emission tomography drugs labeled with positron nuclides such as 18 F-FDG are intravenously administered to a subject, and ⁇ -rays as annihilation radiation in each part of the brain are measured to measure sugar.
- SPECT single photon emission tomography
- nuclides such as 123 I and 99m Tc are used, and the same measurement is performed. The measurement is performed for each cross section of the subject's brain, and a plurality of cross-sectional images as shown in FIG. 1 are generated.
- a cross-sectional image for example, red, yellow, green, blue, from a large region to a small region with a state value (such as a voxel value reflecting a functional state measured by PET, SPECT, etc.) including glucose metabolism and blood flow. Shown with color coding. Diagnosis can be performed by comparing data indicating the state value of the healthy person with data indicating the state value of the subject. At this time, using a computer, the external shape of the cross-sectional image of the healthy person and the cross-sectional image of the subject are matched by some method, the Z score at each part is calculated, and this difference image is presented.
- a state value such as a voxel value reflecting a functional state measured by PET, SPECT, etc.
- Diagnosis can be performed by comparing data indicating the state value of the healthy person with data indicating the state value of the subject.
- the external shape of the cross-sectional image of the healthy person and the cross-sectional image of the subject are matched
- the data indicating the state value of the healthy person is an average value and a standard deviation of the data at each site where a plurality of healthy persons are measured, and is called a normal brain database.
- the Z score is obtained by dividing the difference between the state value of the subject at each part and the state value of the normal brain database by the standard deviation of the normal brain database at the part (see the following formula (1)).
- Z (x, y, z) (I mean (x, y, z) -I (x, y, z)) / SD (x, y, z) Hz
- Z (x, y, z) is the Z score at the coordinate position x, y, z
- I mean (x, y, z) is the state value (PET or SPECT) of the healthy person at the coordinate point.
- I (x, y, z) is the state value of the subject at that coordinate point
- SD (x, y, z) is at that coordinate point This is the standard deviation of the state value of healthy people.
- I mean (x, y, z) and SD (x, y, z) are obtained from the normal brain database.
- the degree of difference from a healthy person can be clarified by taking into account the standard deviation, not just the difference.
- the largest value among the state values from the brain surface to a part of a predetermined depth is displayed on the brain table as a representative value.
- the Z score is calculated by comparing the data extracted in the brain table thus obtained with the extracted data of the brain table in a healthy person.
- the difference image includes right brain lateral surface RT-LAT, left brain lateral surface LT-LAT, upper surface SUP, lower surface BOTTOM, front surface ANT, rear surface POST, right brain inner surface R-MED, left brain inner surface L- Displayed as MED.
- the upper row is a brain surface blood flow image
- the lower row is a Z score image.
- Such a Z-score image can be used as a basis for determining the severity of the disease in addition to improving the detection ability of the diseased part.
- an average value and a standard deviation of the brain surface extraction state values (representative values) of healthy persons in each brain surface region are prepared as a normal brain database.
- the normal brain database requires brain function images of a plurality of healthy subjects, and is preferably created from at least 10 normal cases in order to obtain highly accurate diagnostic performance (Chen WP et. Al. "Effect of sample size for normal database on diagnostic performance of brain FDG PET for the detection of Alzheimer's disease using automated image analysis "Nucl Med Commun. 2008 Mar; 29 (3): 270-6).
- a subject who performs a brain function image examination in a hospital or the like often has some kind of lesion, so it is not easy to collect image data of a healthy person.
- a normal brain database since the functional image of a healthy person differs depending on the combination of what substance is labeled with which radioisotope, a normal brain database must be generated for each combination suitable for disease diagnosis. It was. Such circumstances hindered the creation of a normal brain database.
- an object of the present invention is to provide an apparatus and method that can easily create an alternative normal brain database that can be used as an alternative to a difficult-to-create normal brain database.
- the following are some independent aspects of the invention.
- the alternative normal brain database generation device is a brain region used for determining a brain disease based on a brain function image showing a state of each region of the brain for a plurality of persons including a sick person.
- An apparatus for generating an alternative normal brain database indicating a normal state for each wherein means for normalizing a state value in each part of a brain function image of each person based on a part or all of the state values of the brain function image And a means for spatially deforming the brain function image of each person so as to match the anatomical standard brain, and the brain function image based on the brain function image of each person subjected to normalization and anatomical standardization.
- Each part of the brain used to determine the disease based on the means for excluding the state value estimated to show brain disease, and the brain function image excluding the state value estimated to show brain disease, Each part of the brain used for determining the brain disease And means for calculating at least an average state value and generating an alternative normal brain database indicating the normal state, wherein the excluding means includes a deviation / standard deviation of a plurality of state values for each region.
- the maximum value is calculated as ESD (extream studentized deviate), and when the ESD exceeds a critical value ⁇ calculated based on a predetermined risk ⁇ , at least the state value selected as the ESD is excluded. It is characterized by that.
- the exclusion means is for each part, i) The state value group of all the multiple people is the target state value group, ii) Calculate the ESD for the target state value group and record it as a departure state value; iii) Determine whether the ESD exceeds the critical value ⁇ , record the determination result in association with the deviation state value, iv) If the repetition has not reached the predetermined number of times, remove the departure state value from the target state value group to form a new target state value group, and execute the above ii) and the following: v) If the predetermined number of repetitions is reached, the departure state values before the departure state value when the ESD finally exceeds the critical value ⁇ are all excluded.
- the alternative normal brain database generation device is characterized in that the risk ⁇ is set in the range of 0.1 to 0.6.
- the alternative normal brain database generation device is characterized in that the risk ⁇ is set in the range of 0.2 to 0.5.
- the alternative normal brain database generation device is characterized in that the risk ⁇ is set in the range of 0.2 to 0.4.
- the alternative normal brain database generation device is based on the brain function images of each person that has been normalized and anatomically standardized, and is representative of state values vertically from the brain surface to a predetermined depth.
- an alternative normal brain database for making a determination based on the brain surface state image can be constructed.
- a brain surface blood flow image and a brain surface sugar metabolism image can be generated using a brain blood flow image and a sugar metabolism image as the brain function image.
- the brain function image is a cerebral blood flow image indicating a blood flow corresponding value (a voxel value reflecting a blood flow measured by PET, SPECT, or the like) as a state value. It is characterized by being.
- the generation means uses a brain function that is used when determining the brain disease based on the brain function image excluding the state value estimated to indicate brain disease. For each part, an average state value and a standard deviation are calculated, and a database indicating the normal state is generated.
- the alternative normal brain surface database generation device is an alternative used when determining a brain disease based on a cerebral blood flow image indicating a blood flow state of each part of the brain for a plurality of persons including the sick
- a device for generating a normal brain surface database which normalizes cerebral blood flow correspondence values in each part of each person's cerebral blood flow image based on part or all of the cerebral blood flow correspondence values Based on the cerebral blood flow image of each person that has been normalized and anatomically standardized, and a means for spatially deforming each person's cerebral blood flow image to match the anatomical standard brain
- An alternative normal brain database generation device uses a brain used to determine a brain disease based on a brain function image showing the state of each part of the brain for a plurality of persons including a sick person.
- An apparatus for generating an alternative normal brain database indicating a normal state for each part of the brain, wherein the state values in each part of the brain function image of each person are normalized based on a part or all of the state values of the brain function image Based on the brain function image of each person that has been normalized and anatomically standardized, means for spatially deforming each person's brain function image to match the anatomical standard brain ,
- For each part of the brain used for determining the brain disease means for excluding a state value estimated to show brain disease, and a brain function image excluding the state value estimated to show brain disease Used to determine the brain disease based on For each part of the brain, and it means for calculating an average of at least state value to generate an alternate normal brain database indicating the normal state.
- the excluding means excludes outliers obtained at each site based on a statistical technique as state values estimated to indicate brain disease.
- the brain function image is a cerebral blood flow image showing a blood flow corresponding value.
- An alternative normal brain database generation device is based on a cerebral blood flow image of each person subjected to normalization and anatomical standardization, from a brain surface to a predetermined depth perpendicular to the inside.
- a means for selecting a representative value of the cerebral blood flow correspondence value and generating a cerebral blood flow image of each person as a blood flow correspondence value of the brain surface; Are used as each part of the brain used in determining the brain disease.
- normalization means corresponds to step S13 in FIG.
- the “anatomical deformation means” means at least a function of deforming a brain function image of a subject from an anatomical viewpoint, and in the embodiment, step S2 in FIG. 6 corresponds to this.
- Disease data exclusion means means a function having at least a function of excluding a state value of a site considered to be a disease, and also having a function of excluding abnormal values generated by processing such as normalization. May be. In the embodiment, step S36 in FIG. 21 and step S42 in FIG. 22 correspond to this.
- the “database generating means” means a function having a function of generating a database necessary for disease determination based on at least a plurality of brain function images.
- step S37 in FIG. 21 or step in FIG. S43 corresponds to this.
- Program is a concept that includes not only a program that can be directly executed by the CPU but also a source-format program, a compressed program, an encrypted program, and the like.
- FIG. 1 is a diagram showing a cerebral blood flow cross-sectional image.
- FIG. 2 is a diagram showing a brain surface blood flow image.
- FIG. 3 is a block diagram of an alternative normal brain database generation device according to an embodiment of the present invention.
- FIG. 4 is a block diagram of an alternative normal brain database generation device according to another embodiment.
- FIG. 5 is a hardware configuration of the alternative normal brain database generation device.
- FIG. 6 is a flowchart of the alternative normal brain database generation program recorded on the hard disk.
- FIG. 7 is a detailed flowchart of cerebral blood flow data acquisition (step S1 in FIG. 6).
- FIG. 8 shows normalized cerebral blood flow data.
- FIG. 9 is a diagram illustrating the X axis, the Y axis, and the Z axis related to the head.
- 10 and 11 are detailed flowcharts of the standardization of the blood flow image (step S2 in FIG. 6). 10 and 11 are detailed flowcharts of the standardization of the blood flow image (step S2 in FIG. 6). 12 and 13 are detailed flowcharts for identifying the median sagittal section in the blood flow image (step S22 in FIG. 10). 12 and 13 are detailed flowcharts for identifying the median sagittal section in the blood flow image (step S22 in FIG. 10).
- FIG. 14 is a diagram for explaining probabilistic sign change (SSC).
- FIG. 15 is a detailed flowchart for specifying the AC-PC line and its midpoint (step S23 in FIG. 10).
- FIG. 16 is a diagram for explaining identification of an AC-PC line.
- FIG. 17 is a diagram illustrating estimation of the landmark TH for specifying the AC-PC line.
- FIG. 18 is a diagram showing a linear transformation for matching a blood flow image to an anatomical standard brain.
- FIG. 19 is a diagram illustrating nonlinear transformation for matching a blood flow image to an anatomical standard brain.
- FIG. 20 is a detailed flowchart of generation of a brain surface blood flow image (step S27 in FIG. 11).
- 21 and 22 are flowcharts for generating an alternative normal brain database.
- 21 and 22 are flowcharts for generating an alternative normal brain database.
- FIG. 23 is a detailed flowchart of processing for rejecting a blood flow corresponding value estimated to be a diseased part for a target voxel.
- FIG. 23 is a detailed flowchart of processing for rejecting a blood flow corresponding value estimated to be a diseased part for a target voxel.
- FIG. 24 is a detailed flowchart of processing for rejecting a blood flow corresponding value estimated to be a diseased part for a target voxel. It is a table
- FIG. 27 shows a part of the generated cerebral blood flow standard database.
- FIG. 28 shows a part of the generated brain surface blood flow standard database.
- FIG. 29 is a detailed flowchart of a rejection process according to another embodiment.
- FIG. 30 is a table of T for the rejection process.
- FIG. 31 shows an image (brain surface extracted image) (a) average value image and (b) standard deviation image configured using a virtual image database.
- FIG. 32 shows an image (brain surface extracted image) (a) average value image and (b) standard deviation image constructed using data obtained by adding random data with an average value of 80% and a variation coefficient of 40% to the virtual image database. is there.
- A Average value image
- FIG. 3 shows a configuration of an alternative normal brain database generation device according to an embodiment of the present invention.
- This apparatus can generate an alternative normal brain database based on brain function images of a plurality of persons including a sick person. All of the brain function images given to this device may be images of a sick person.
- the anatomical standardization means 2 takes in the brain function image of each person and performs a spatial deformation so as to match the anatomical standard brain (for example, the Tarailach standard brain). Thereby, it is possible to eliminate the morphological difference of each person's brain and standardize it.
- the anatomical standard brain for example, the Tarailach standard brain
- the disease data exclusion means 4 identifies a state value estimated to indicate a brain disease by comparing the state value of each person for each part of the standardized brain function image, and performs a process of excluding this. Do. Such an exclusion process can be performed based on a statistical method.
- the disease data exclusion means 4 performs the above exclusion process for all parts.
- the database generation means 6 includes an average value calculation means 8 and a standard deviation calculation means 10.
- the average value calculation means 8 calculates an average value for each part based on the brain function image from which the disease data is excluded.
- the standard deviation calculation means 10 calculates the standard deviation for each part based on the brain function image from which the disease data is excluded. These average values and standard deviations are associated with information indicating the position of each part and recorded as a substitute normal brain database.
- FIG. 4 shows a configuration of an alternative normal brain database generation device according to another embodiment. This device generates an alternative normal brain database indicating the state of the brain surface for use in 3D-SSP and the like.
- the anatomical standardization means 2 takes in the brain function image of each person and performs spatial deformation so as to match the anatomical standard brain.
- the brain surface state image generating means 3 selects a state value representative of the brain surface from state values in the vicinity of the brain surface, and sets this as the state value of the brain surface. This is performed for all the parts, and each person's brain surface state image (also called a brain surface functional image) is obtained.
- the disease data excluding means 4 is a process for identifying and excluding a state value estimated to indicate brain disease by comparing each person's state value for each part of the standardized brain surface state image. I do.
- the database generation means 6 includes an average value calculation means 8 and a standard deviation calculation means 10.
- the average value calculation means 8 calculates an average value for each part of the brain surface based on the brain surface state image from which the disease data is excluded.
- the standard deviation calculating means 10 calculates the standard deviation for each part based on the brain surface state image from which the disease data is excluded. These average values and standard deviations are associated with information indicating the position of each part and recorded as a substitute normal brain database.
- FIG. 5 shows a hardware configuration for realizing the alternative normal brain database generation device of FIGS. 3 and 4.
- an apparatus that simultaneously realizes the functions of FIG. 3 and FIG. 4 is shown.
- an apparatus that realizes only one of the functions can be used.
- cerebral blood flow corresponding values are shown as an example of brain state values.
- a hard disk 14, a CD-ROM drive 16, a display 18, a memory 20, and a keyboard / mouse 22 are connected to the CPU 12.
- the display 18 is for displaying a brain image or the like.
- the memory 20 is used as a work area for the CPU 12.
- the keyboard / mouse 22 is used by an operator of the apparatus to input commands and the like.
- the hard disk 14 stores an operating system 24 such as WINDOWS (trademark), an alternative normal brain database generation program 26, standard brain data 28, and the like. These programs and data are installed from a recording medium such as the CD-ROM 30 via the CD-ROM drive 16.
- an operating system 24 such as WINDOWS (trademark)
- an alternative normal brain database generation program 26 standard brain data 28, and the like.
- These programs and data are installed from a recording medium such as the CD-ROM 30 via the CD-ROM drive 16.
- the alternative normal brain database generation program 26 performs its function in cooperation with the operating system 24.
- FIG. 6 shows a flowchart of an alternative normal brain database generation program. This program has three steps: acquisition and normalization of cerebral blood flow data (step S1), anatomical standardization (step S2), and generation of an alternative normal brain database (step S3).
- step S1 acquisition and normalization of cerebral blood flow data
- step S2 anatomical standardization
- step S3 generation of an alternative normal brain database
- the cerebral blood flow correspondence value is described as an example, but the present invention can also be applied to other functional images.
- the CPU 12 acquires SPECT data of a plurality of subjects including sick persons (step S11). If the apparatus shown in FIG. 5 is connected to the SPECT apparatus via a LAN or the like, it can be obtained by directly receiving data. Further, the data recorded on the recording medium by the SPECT apparatus can be read from the recording medium and acquired.
- the SPECT data is constituted by projection data acquired from the subject by the SPECT apparatus.
- the CPU 12 reconstructs, for each person, a three-dimensional image composed of voxels having a vertical and horizontal height of about 2 mm based on the SPECT data (step S12).
- the absolute value of the voxel value in the three-dimensional image varies depending on the measurement conditions of the measuring device and the sample. Therefore, the CPU 12 calculates, for each person's respective voxel value, the blood flow corresponding value of each part of the three-dimensional image as an average blood flow corresponding value (whole brain average value of voxel values) in the entire person's brain.
- the blood flow corresponding value data is normalized and recorded in the hard disk 14 (step S13). Normalization can absorb fluctuations due to measurement conditions and the like.
- normalization is performed based on the mean blood flow corresponding value for the whole brain, but normalization may be performed based on the mean blood flow corresponding value for the thalamus, cerebellum, bridge, and sensorimotor area.
- the thalamus, cerebellum, bridge, sensorimotor area, etc. in the SPECT data can be specified by superimposing the subject's image on the anatomical standard brain in which these parts are specified in advance.
- the site used as a standard for normalization is preferably a site where a decrease in blood flow corresponding value is not observed in a disease example to be diagnosed.
- FIG. 8 shows the data of the normalized three-dimensional blood flow image generated in this way.
- X, Y, and Z represent coordinate positions, where X is the left-right direction toward the head, Y is the front-back direction of the head, and Z is the up-down direction of the head.
- FIG. 9 shows the X axis, the Y axis, and the Z axis.
- the CPU 12 spatially transforms each person's normalized three-dimensional blood flow image to match the anatomical standard brain (step S2 in FIG. 6).
- the CPU 12 reads the first subject's normalized three-dimensional blood flow image (see FIG. 8) from the hard disk 14 (step S21).
- the CPU 12 determines the median sagittal section of the read normalized three-dimensional blood flow image (step S22).
- the median sagittal section is a YZ plane in a line ⁇ passing through the left and right center of the head in FIG. 9. In other words, it is a surface that divides the head equally to the left and right.
- the determination of the median sagittal section is generally performed as follows. First, a center point of an image is determined, and a mid-sagittal section is assumed assuming a YZ plane passing through the center point. Next, the assumed mid-sagittal section is used as a folding surface, and the right image is folded to the left to generate a symmetrical image. The degree of similarity between the folded three-dimensional blood flow image and the original three-dimensional blood flow image is determined. The hypothetical mid-sagittal section is moved in the X direction, rotated about the Z axis and the Y axis, and the similarity is determined for each.
- the similarity between the folded 3D blood flow image and the original 3D blood flow image should be maximized, and the folded surface is the mid-sagittal section. Therefore, in this embodiment, the surface having the highest similarity between the two images is determined as the mid-sagittal section.
- the CPU 12 determines simple center points x 0 , y 0 , z 0 based on the coordinate positions of the three-dimensional blood flow image. A point moved by ⁇ x with respect to the center points x 0 , y 0 , z 0 is set as the coordinate center (step S222).
- this surface is rotated by ⁇ z around the Z axis and rotated by ⁇ y around the Y axis.
- the rotated YZ plane is assumed to be a median sagittal section (step S225).
- the CPU 12 folds the image on the right side of the hypothetical mid-sagittal section to generate a symmetrical folded three-dimensional blood flow image (step S226).
- the similarity between the original three-dimensional blood flow image and the folded three-dimensional blood flow image is calculated (step S227).
- the similarity is obtained by calculating an SSC (Stochastic sign change) value.
- the concept of similarity calculation by SSC is as follows.
- the degree should be high. Therefore, in SSC, the number of changes in the sign of the difference between adjacent pixels is calculated as an SSC value, and the higher the SSC value, the higher the similarity.
- the ZX 1 to ZX N planes are assumed, and further the YX 1 to YX N planes are assumed.
- the SSC value when scanning in the direction is calculated.
- the SSC values calculated for each of these surfaces are summed to obtain the SSCx value.
- the SSCy value related to the Y direction is the SSCy value obtained by adding all the SSC values when scanning is performed in the Y direction for each of the ZY 1 to ZY N planes and the YX 1 to YX N planes.
- the SSCz value related to the Z direction is the SSCz value obtained by adding all the SSC values when scanning is performed in the Z direction for each of the ZY 1 to ZY N planes and the ZX 1 to ZX N planes.
- the CPU 12 adds the above SSCx, SSCy, and SSCz to obtain the SSC value ( ⁇ X, ⁇ z, ⁇ y) regarding the hypothetical mid-sagittal section and stores it in the memory 20.
- the CPU 12 changes ⁇ x for each voxel from ⁇ I to I to move the coordinate center, and changes ⁇ z and ⁇ y by a predetermined angle from ⁇ I to ⁇ I , and all these combinations.
- SSC values ( ⁇ X, ⁇ z, ⁇ y) are calculated by assuming a median sagittal section (steps S221, S223, S224).
- the CPU 12 selects ⁇ X, ⁇ z, ⁇ y that can obtain the largest SSC values ( ⁇ X, ⁇ z, ⁇ y). Based on this, the median sagittal section is determined (step S229).
- the median sagittal section is determined (step S229). For the method of determining the median sagittal section, refer to Minoshima et al "An Automaeted Method for Rotational Correction and Centering of Three-Dimensional Functional Brain Images" J Nucl Med 1992; 33: 1579-1585. take in.
- the AC-PC line and midpoint are then determined (step S23 in FIG. 10).
- the frontal lobe pole (OP), hypothalamus (TH), underside of the corpus callosum (CC), frontal lobe pole (FP) in the median sagittal section are determined, and these four landmarks are connected as straight lines. Determine the AC-PC line.
- FIG. 15 shows a process for determining the AC-PC line and its midpoint.
- CPU12 reads the blood-flow cross-sectional image of a median sagittal cross section. Moreover, the blood flow image in the cross section of several front and back surfaces horizontal to the median sagittal cross section is read (step S231). An example of a blood flow image in the mid-sagittal section is shown in FIG. 16A.
- CPU 12 detects the external outline of the brain and the outline of the boundary between white matter and gray matter for each of the read blood flow images (step S232). These contours can be extracted from the blood flow image depending on the presence or absence of blood flow. An example of the detected contour image is shown in FIG. 16B.
- the rearmost end of the contour is identified and used as the occipital lobe pole (OP) (see FIG. 16B).
- the occipital lobe pole (OP) is determined for all of the plurality of cross sections, and the average value of the YZ coordinates is taken to determine the position of the occipital lobe pole (OP) on the median sagittal section (step S233).
- the CPU 12 searches for a U-shaped contour portion as shown in FIG. 16C.
- the point where the tangent drawn from the occipital lobe pole (OP) to the U-shaped part contacts the U-shaped part is the lower surface (CC) of the frontal corpus callosum.
- the lower surface (CC) of the front part of the corpus callosum is determined for all of the plurality of cross sections, and the average value of the YZ coordinates is taken to determine the position of the lower surface (CC) of the front part of the corpus callosum on the median sagittal section (step S234). ).
- the CPU 12 specifies the most advanced part of the contour and sets it as the frontal lobe pole (FP) (see FIG. 16B).
- the frontal lobe pole (FP) is determined for all of the plurality of cross sections, and the average value of the YZ coordinates is taken to determine the position of the frontal lobe pole (FP) on the median sagittal cross section (step S235).
- the CPU 12 determines the hypothalamus (TH) (step S236).
- the CPU 12 first searches for the center point of the thalamus.
- the center point of the thalamus can be found as a point in the vicinity of the AC-PC line and the blood flow corresponding value shows a maximum value.
- the most within the predetermined radius from the midpoint of the line connecting the frontal lobe pole (FP) and the occipital lobe pole (OP) (in this embodiment, the length that is 1/10 of the length from FP to OP) A part having a large blood flow correspondence value is found and used as the center point of the thalamus.
- the CPU 12 draws a circle with a predetermined radius r from the center point.
- a point where the circle and a tangent line from the occipital lobe pole (OP) are in contact is P1
- P2 a point on the opposite side of P1
- the CPU 12 changes the radius r from 1 pixel to 12 pixels in increments of 0.5 pixels, and records the change in the blood flow corresponding value at the point P2 at each r in the memory 20.
- FIG. 17B shows changes in blood flow corresponding values plotted in this way.
- r increases from 0 in the center toward the left side.
- the blood flow corresponding value at the point P2 decreases and finally reaches the minimum value. This minimum value corresponds to the lateral ventricular (LV).
- LV lateral ventricular
- the CPU 12 specifies a radius r indicating a blood flow correspondence value of 70% when the maximum blood flow correspondence value is 100% and the minimum blood flow correspondence value is 0%. This is determined as the end of the thalamus. Therefore, the point P1 at the radius r can be obtained as the hypothalamus (TH).
- the CPU 12 obtains the hypothalamus (TH) for all of the plurality of cross sections, takes the average value of the YZ coordinates, and determines the position of the hypothalamus (TH) on the median sagittal cross section (see FIG. 16B).
- the CPU 12 calculates a regression line connecting the occipital lobe pole (OP), hypothalamus (TH), underside of the corpus callosum (CC), and frontal lobe pole (FP) determined in the mid-sagittal section, Is an AC-PC line ⁇ (step S237). Subsequently, the midpoint ⁇ is calculated (step S238) (see FIG. 16D).
- AC-PC line and midpoint determination method is incorporated into this specification with reference to Minoshima et al "Autometed Detection of the Intercommissural Line for Stereotactic Localization of Functional Brain Images" J Nucl Med 1993; 34: 322-329 .
- the CPU 12 next records the subject's three-dimensional cerebral blood flow image on the hard disk 14 in advance. It is superimposed on the standard brain image data 28 that has been stored (step S24 in FIG. 10). At this time, the CPU 12 matches the direction by matching the AC-PC line of the three-dimensional cerebral blood flow image of the subject with the AC-PC line of the standard brain image, and the AC-PC of the three-dimensional cerebral blood flow image of the subject. By matching the midpoint of the PC line with the midpoint of the AC-PC line of the standard brain image data 28, the front and back positions are matched.
- an 18 F-FDGPET image is used as the standard brain image data 28. This image is obtained by averaging three-dimensional images of many subjects and expressing them as an anatomical standard brain shape.
- the standard brain image of the 18 F-FDGPET image is generally used and can be easily obtained.
- the CPU 12 generates a line perpendicular to the AC-PC line on the mid-sagittal cross section passing through the Y-axis of the AC-PC line of the subject's 3D cerebral blood flow image and the midpoint of the AC-PC line.
- the normal of the median sagittal section passing through the midpoint of the Z axis and the AC-PC line is taken as the X axis. Therefore, the midpoint ⁇ of the AC-PC line is the coordinate origin.
- the CPU 12 acquires the Y coordinate of the frontmost point (ie, FP) and the rearmost point (ie, OP) in the subject's three-dimensional cerebral blood flow image, and the uppermost point (point T in FIG. 18).
- each coordinate value may be recorded in advance.
- the CPU 12 deforms the subject's 3D cerebral blood flow image so that the external contour of the subject's 3D cerebral blood flow image matches the external contour of the standard brain image.
- a three-dimensional cerebral blood flow image (shown by a solid line) of a subject is superimposed on a standard brain image (shown by a two-dot chain line).
- the CPU 12 finds the uppermost point T in the outline of the three-dimensional cerebral blood flow image, and acquires the Z coordinate Lz.
- the uppermost point T ′ in the contour of the standard brain image is found, and the Z coordinate L′ z is obtained.
- the CPU 12 transforms the Z coordinate value of the subject's three-dimensional cerebral blood flow image according to the following equation.
- Z ′ Z ⁇ (L Z / L ′ Z )
- Z ′ Z coordinate value of the three-dimensional cerebral blood flow image of the subject after deformation
- Z is the Z coordinate value of the original three-dimensional cerebral blood flow image of the subject
- Lz is the Z coordinate value of the point T
- L ′ is the Z coordinate value of the point T ′.
- the upward image of the subject's three-dimensional cerebral blood flow image can be matched with the brain standard image.
- the downward direction, the right direction, the left direction, the forward direction, and the backward direction are also deformed so as to match the brain standard image.
- FIG. 19A In the standard brain image, as shown in FIG. 19A, some landmarks C are defined in white matter. Further, a landmark S on the brain surface is defined in correspondence with each white matter landmark C. In FIG. 19A, brain surface landmarks S1, S2, and S3 are defined in association with the white matter landmark C1.
- the CPU 12 reads the blood flow corresponding value from the three-dimensional cerebral blood flow image of the subject along the straight line L1 connecting the landmarks C1 and S1, and generates a profile curve thereof (solid line in FIG. 19B). On the other hand, the profile curve on the straight line L1 in the standard brain image is recorded in advance (two-dot chain line in FIG. 19B).
- the CPU 12 fixes the landmark C1 along the direction of the straight line L1 so that the subject's profile curve on the straight line L1 approaches the profile curve of the standard brain image, and the subject's three-dimensional Stretch the cerebral blood flow image. Similar processing is performed for the straight lines L2 and L3. The region between the straight lines L1 and L2 (L2 and L3) is deformed in consideration of both the deformation rate on the straight line L1 and the deformation rate on the straight line L2. In the above description, the non-linear transformation on the two-dimensional plane is shown for ease of explanation, but the transformation is applied to the three-dimensional space.
- a brain surface blood flow image is then generated (step S27 in FIG. 11). Details of generation of the brain surface blood flow image are shown in FIG.
- the CPU 12 identifies each voxel on the brain surface (the surface of the brain) of the standardized three-dimensional cerebral blood flow image of the subject and assigns an identification number (step S271).
- the number of voxels on the brain surface is about 20,000.
- the CPU 12 performs the following processing on the voxels of each brain surface.
- a blood flow correspondence value of 6 voxels is acquired from the voxel perpendicular to the internal direction with respect to the brain surface (step S273).
- the largest value among these blood flow correspondence values is selected as a representative value, and is set as the blood flow correspondence value of the voxel of the brain table (step S274).
- This process is performed for all voxels on the brain surface. As a result, it is possible to generate a brain surface blood flow image in which a representative value of the blood flow corresponding value is shown in the brain surface.
- the CPU 12 records the cerebral blood flow image generated in step S26 and the cerebral blood flow image generated in step S27 on the hard disk 14 (step S28 in FIG. 11).
- step S30 If there is data of the next subject, the data is read (step S30), and a cerebral blood flow image and a cerebral blood flow image are generated in the same manner as described above. If the cerebral blood flow image and the cerebral blood flow image are generated for the data of all the subjects (step S29), the process ends.
- Step S3 the CPU 12 next generates an alternative normal brain database ( Step S3 in FIG. If the subject is a healthy person, a normal brain database can be generated by obtaining an average value and a standard deviation for each part of the standardized brain blood flow image and brain surface blood flow image of each subject. However, in this embodiment, some or all of the subjects are sick. Therefore, it is necessary to generate an alternative normal brain database by excluding blood flow corresponding values of sites determined to be diseases.
- FIGS. CPU12 extracts the cerebral blood flow corresponding value in a certain voxel (target voxel) from the cerebral blood flow images of all subjects (step S34).
- FIG. 25 shows the extracted blood flow correspondence values.
- the patient ID is an identifier assigned to each subject.
- the blood flow correspondence value is normalized.
- step S36 the CPU 12 rejects outliers.
- outliers are rejected by a GESD (Generized extream studentized deviate) test.
- the detailed flowchart is shown in FIGS.
- the CPU 12 sets the number of subjects (that is, the number of blood flow corresponding values in the target voxel) to n, sets the repetition coefficient i to 0, and sets the maximum rejection number TN to a predetermined value (step S361).
- the maximum rejection number TN is the maximum possible number of data to be rejected, and can be set arbitrarily.
- a value obtained by rounding down the fractional number of n / 2 to an integer is calculated as TN and used.
- the CPU 12 calculates the arithmetic mean of the blood flow corresponding values in the target voxel, excluding the blood flow corresponding values that have already deviated (step S362), and calculates the standard deviation (step S363). Here, yet since there is no a departure, it calculates the arithmetic MEAN and a standard deviation SD of all bloodstream associated value I 1 of the subject, I 2 ... I n.
- the CPU 12 calculates deviations D 1 , D 2 ... D n of each subject from the MEAN, and calculates R 1 , R 2 ... R n for each subject based on the following equation.
- R Deviation D / Standard deviation SD
- the CPU 12 selects the maximum one of the calculated R 1 , R 2 ... R n of each subject and sets it as R max1 (step S364). This R max1 is called ESD (extream studentized deviate).
- the CPU 12 records the ID and R of the subject selected as R max1 in the memory 20 as deviation data (only the subject ID may be recorded).
- the CPU 12 calculates a probability P based on the following equation (step S365).
- ⁇ is a degree of danger, and is a degree that determines how much rejection is performed.
- the risk degree ⁇ takes a value of 0 to 1, and the greater the value, the greater the degree of rejection. Therefore, an outlier can be reliably rejected as the degree of risk ⁇ increases, but on the other hand, the risk of rejecting a normal value by mistake increases. Conversely, the smaller the degree of risk ⁇ , the lower the possibility that the normal value will be erroneously rejected, but on the other hand, the possibility that the outlier cannot be rejected increases.
- the value of the risk ⁇ an appropriate value is set in advance (recorded on the hard disk 14) according to the target disease and the type of functional image, and this is used.
- the risk ⁇ is preferably 0.1 to 0.6, and preferably 0.2 to 0.5. More preferably, it is 0.2 to 0.4 (see examples described later).
- the CPU 12 obtains a t value (T) of ni ⁇ 1 + 1 degrees of freedom corresponding to the probability P from the T distribution table (step S366).
- T a t value of ni ⁇ 1 + 1 degrees of freedom corresponding to the probability P from the T distribution table.
- the critical value ⁇ is calculated by the following equation (step S367).
- the CPU 12 determines whether or not R max1 which is the ESD calculated in step S364 is larger than the critical value ⁇ (step S368). If it is larger, a flag indicating that it should be rejected is recorded for the deviation data R max1 recorded in the memory 20 (step S369). For example, when R 8 is selected as R max1 in step S364 and is larger than the critical value ⁇ , a rejection flag is recorded as shown in the first line of FIG. When R max1 is equal to or smaller than the critical value ⁇ , no rejection flag is recorded.
- step S370 the CPU 12 increments i (step S370) and executes step S362 and subsequent steps.
- i 1 and step S362 and subsequent steps are executed.
- step S362 and S363 the arithmetic mean MEAN and the standard deviation SD are calculated except for the deviation data recorded in the memory 20.
- step S364 R max2 is calculated.
- the CPU 12 determines whether or not R max2 is larger than the critical value ⁇ (step S368), and if so, records a rejection flag in the memory 20 for the deviation data R max2 as described above (step S369). .
- step S370 increments i (step S370) and executes step S362 and subsequent steps.
- the CPU 12 repeats this until i reaches the maximum rejection number TN (step S371).
- the CPU 12 searches the deviation table for data in which the rejection flag is recorded last. That is, the data with the largest i is specified among the data in which the rejection flag is recorded.
- CPU12 makes all the data (data with small i) before the data with which the rejection flag was recorded last as rejection object (step S372).
- the CPU 12 calculates and records an average value and a standard deviation for the blood flow corresponding values remaining without being rejected (FIG. 21, step S37). In this way, the average value and standard deviation of the blood flow corresponding values in the voxels (11, 25, 135) are calculated and recorded (see FIG. 24).
- the CPU 12 calculates and records the average value and the standard deviation for all the voxels as described above (steps S31, S32, S33, S38). In this way, an alternative normal brain database for cerebral blood flow images as shown in FIG. 24 is completed.
- the CPU 12 calculates an average value and a standard deviation for each voxel of the brain surface in the same manner for the brain surface blood flow image.
- CPU12 extracts the cerebral blood flow corresponding value in the voxel of a certain brain surface from the brain surface blood flow image of all the subjects (step S40).
- the CPU 12 executes the above-mentioned GESD test (FIGS. 23 and 24) and rejects the outlier (step S42).
- the CPU 12 calculates and records an average value and a standard deviation for the blood flow corresponding values that remain without being rejected (step S43). In this way, the average value and standard deviation of blood flow corresponding values in the voxels of the brain surface are calculated and recorded (see FIG. 28).
- the CPU 12 calculates and records the average value and the standard deviation for all brain surface voxels as described above (steps S39 and S44). In this way, an alternative normal brain database for the brain surface blood flow image as shown in FIG. 28 is completed.
- outliers are rejected using the GESD test in order to reject blood flow corresponding values estimated to indicate brain disease.
- the rejection may be performed using other statistical methods.
- the CPU 12 sets the blood flow corresponding value Ip (x, y, z) of a subject as a test target (step S352). Subsequently, an average value I mean and standard deviation SD of blood flow corresponding values I (x, y, z) of all subjects are calculated (step S353). Subsequently, the absolute value obtained by subtracting the average value I mean from the blood flow corresponding value Ip of the test subject is divided by the standard deviation SD to calculate T 0 (step S354).
- the limit value T alpha and T 0 are determined in advance (step S355).
- the limit value T ⁇ is determined from the number M of subjects and the risk factor ⁇ .
- the CPU 12 obtains a limit value T ⁇ based on the number M of subjects and the risk factor ⁇ given by the user according to the table shown in FIG. 30 (recorded in advance on the hard disk).
- the risk factor ⁇ is a factor indicating how much rejection is performed, and the larger the risk factor ⁇ , the more data is rejected.
- the limit value T alpha becomes 1.93 from the table of FIG. 30.
- the CPU 12 repeats the above processing using the blood flow corresponding values of all subjects as the test target (steps S351 and S357).
- outliers having a large blood flow correspondence value and outliers having a small blood flow correspondence value are excluded at the same risk factor ⁇ .
- the risk rate on the side that is more likely to show brain disease (for example, the small blood flow value side) is set to ⁇ 1
- the risk rate on the opposite side (for example, the large blood flow value side) is set to be smaller than ⁇ 1.
- the number M of subjects must be at least 4 or more.
- the cerebral blood flow image by SPECT or PET has been described, but it can also be applied to other functional images.
- spatial deformation anatomical standardization
- normalization may be performed after performing spatial deformation.
- normalization and spatial deformation may be performed simultaneously.
- the normal normal database created using a plurality of head SPECT images was used to verify the alternative normal brain database creation method according to the present invention.
- original image data surface pixel data (hereinafter referred to as original image data) of ChibaDB_ver2IMP60-64AGLBMNSFM (supplied by Chiba University), which is a normal database created based on IMP-administered head SPECT data of healthy subjects It was.
- Random data with a normal distribution with a variation coefficient of 10% is calculated for each pixel with respect to the average value of each pixel of the original image data, and a virtual image database consisting of 40 sets of virtual image data is created (images are This is shown in Fig. 31. Hereinafter, it is simply referred to as a virtual image database.
- Ten random data generated with the average values and coefficient of variation described in Tables 1 to 4 were mixed into the data set of each pixel in this virtual image database to create a verification data group (50 sets of image data). Equivalent).
- the t-test (significance level 5%) and F test with the virtual image database are performed as they are without performing the GESD test on the data group for each pixel for verification, and The presence or absence of a significant difference was confirmed for each pixel (Comparative Examples 1 to 4).
- the number of pixels for which the standard deviation and the average value are determined to have a significant difference from the virtual image database fluctuated depending on the risk ⁇ in the GESD test. That is, the number of pixels for which the standard deviation and the average value are determined to be significantly different from the virtual image database is further decreased at the risk level ⁇ 0.1 to 0.6, and at the risk level ⁇ 0.2 to 0.5. It further decreased and decreased particularly at the risk level ⁇ of 0.2 to 0.4. From this result, the risk ⁇ in the GESD test is preferably 0.1 to 0.6, more preferably 0.2 to 0.5, and more preferably 0.2 to 0.4. It was confirmed that it was particularly preferable.
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Epidemiology (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Primary Health Care (AREA)
- Molecular Biology (AREA)
- Heart & Thoracic Surgery (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Optics & Photonics (AREA)
- High Energy & Nuclear Physics (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Dentistry (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Neurology (AREA)
- Neurosurgery (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
Abstract
Description
診断は、健常者の状態値を示すデータと、被験者の状態値を示すデータとを比較することによって行うことができる。この際、コンピュータを用いて、健常者の断面画像と被験者の断面画像の外形を何らかの手法で一致させ、各部位におけるZスコアを算出し、この差分画像を提示することが行われている。このような方法を可能とする手法としては、ワシントン大学の蓑島によって開発された3-Dimentional stereotaxic surface projection(3D-SSP)やロンドンのハマースミス病院のFristonらによって開発されたStatistical Parametoric Mapping(SPM)がよく知られている。
ここで、Z(x,y,z)は、座標位置x,y,zにおけるZスコアであり、Imean(x,y,z)は当該座標点における健常者の状態値(PETやSPECTによって計測された機能状態を反映したボクセル値)の平均値であり、I(x,y,z)は当該座標点における被験者の状態値であり、SD(x,y,z)は当該座標点における健常者の状態値の標準偏差である。Imean(x,y,z)およびSD(x,y,z)は、正常脳データベースから取得する。
(1)この発明に係る代替正常脳データベース生成装置は、疾患者を含む複数人についての脳の各部位の状態を示す脳機能画像に基づいて、脳疾患を判定する際に用いる脳の各部位についての正常状態を示す代替正常脳データベースを生成する装置であって、各人の脳機能画像の各部位における状態値を、脳機能画像の一部又は全部の状態値に基づいて正規化する手段と、各人の脳機能画像を解剖学的標準脳に合致するように空間的な変形を行う手段と、正規化および解剖学的標準化がなされた各人の脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、脳疾患を示すと推定される状態値を除外する手段と、脳疾患を示すと推定される状態値を除外した脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳データベースを生成する手段とを備え、前記除外手段は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴としている。
i)複数人全ての状態値を対象状態値群とし、
ii)対象状態値群についてESDを算出し、逸脱状態値として記録し、
iii)ESDが臨界値λを超えているかどうかを判断し、判断結果を逸脱状態値に対応付けて記録し、
iv)繰り返しが所定回数に達していなければ、対象となっている状態値群から逸脱状態値を取り除き新しい対象状態値群を形成し、上記ii)以下を実行し、
v)所定の繰り返し回数に達すれば、最後にESDが臨界値λを超えた時の逸脱状態値より前の逸脱状態値を全て除外対象とすることを特徴としている。
脳表血流画像生成手段によって生成された各人の脳表血流画像の各部位ごとに、脳疾患を示すと推定される血流対応値を除外する手段と、脳疾患を示すと推定される血流対応値を除外した脳表血流画像に基づいて、前記脳疾患を判定する際に用いる脳表の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳表データベースを生成する手段と、を備えた代替正常脳表データベース生成装置であって、前記除外手段は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴としている。
図3に、この発明の一実施形態による代替正常脳データベース生成装置の構成を示す。この装置は、疾患者を含む複数人の脳機能画像に基づいて代替正常脳データベースを生成することができる。この装置に与えられる脳機能画像は、その全てが疾患者の画像であってもよい。
図5に、図3および図4の代替正常脳データベース生成装置を実現するためのハードウエア構成を示す。なお、以下の実施形態では、図3の機能および図4の機能を同時に実現した装置を示すが、いずれか一方だけの機能を実現した装置とすることもできる。また、以下の実施形態では、脳の状態値の一例として脳血流対応値を示している。
図6に、代替正常脳データベース生成プログラムのフローチャートを示す。このプログラムは、脳血流データの取得と正規化(ステップS1)、解剖学的標準化(ステップS2)、代替正常脳データベースの生成(ステップS3)の3つのステップを備えている。以下の実施形態では、脳血流対応値を例として説明しているが、その他の機能画像についても適用することができる。
脳血流データの取得処理の詳細を、図7に示す。CPU12は、まず、疾病者を含む複数の被験者のSPECTデータを取得する(ステップS11)。図5に示す装置が、SPECT装置とLANなどによって接続されていれば、直接データを受け取ることによって取得することができる。また、SPECT装置によって記録媒体に記録されたデータを、当該記録媒体から読み込んで取得することもできる。SPECTデータは、SPECT装置により被験者から取得された投影データによって構成されている。
次に、CPU12は、各人の正規化三次元血流画像を、解剖学的標準脳に合致するように空間的変形を行う(図6のステップS2)。
次に、CPU12は、読み出した正規化三次元血流画像の正中矢状断面を決定する(ステップS22)。ここで、正中矢状断面とは、図9において、頭の左右中央を通る線αにおける、YZ平面である。つまり、頭を左右均等に分ける面である。
上記のようにして三次元血流画像の正中矢状断面を決定すると、次に、AC-PC線および中点の決定を行う(図10のステップS23)。この実施形態では、正中矢状断面における前頭葉極(OP)、視床下部(TH)、脳梁前部の下面(CC)、前頭葉極(FP)を決定し、この4つのランドマークを結ぶ直線としてAC-PC線を決定する。
以上のようにしてAC-PC線およびその中点が決定されると、CPU12は、次に、被験者の三次元脳血流画像を予めハードディスク14に記録しておいた標準脳画像データ28に重ね合わせる(図10のステップS24)。この際、CPU12は、被験者の三次元脳血流画像のAC-PC線を、標準脳画像のAC-PC線に合致させることで方向を合致させ、被験者の三次元脳血流画像のAC-PC線中点を、標準脳画像データ28のAC-PC線中点に合致させることで前後の位置を合致させる。
被験者の三次元脳血流画像を標準脳画像に重ねて位置合わせした後、CPU12は、標準脳画像に合致するように被験者の三次元脳血流画像に対して、以下に示すような直線的変形を施す。
ここで、Z’は変形後の被験者の三次元脳血流画像のZ座標値、Zは元の被験者の三次元脳血流画像のZ座標値、Lzは点TのZ座標値、L’zは点T’のZ座標値である。
上記のようにして線形変換を行うことにより、被験者の三次元脳血流画像は標準脳画像に概ね合致される。しかし、細部において両画像にはずれがある。CPU12は、標準脳画像のグルコース代謝のプロファイル曲線と、被験者の三次元脳血流画像のプロファイル曲線に基づいて、部分的な変形を行う(ステップS26)。
上記のようにして被験者の三次元脳血流画像を解剖学的に標準化すると、次に、脳表血流画像を生成する(図11のステップS27)。脳表血流画像の生成の詳細を、図20に示す。
以上のようにして解剖学的に標準化された各被験者の脳血流画像および脳表血流画像を生成すると、CPU12は、次に代替正常脳データベースの生成を行う(図6のステップS3)。被験者が健常者であれば、標準化された各被験者の脳血流画像および脳表血流画像の各部位について平均値、標準偏差を求めることで正常脳データベースを生成することができる。しかし、この実施形態においては、被験者の一部又は全部が疾患者である。そこで、疾患と判断される部位の血流対応値を除外して代替正常脳データベースを生成する必要がある。
CPU12は、算出した各被験者のR1,R2...Rnの中から最大のものを選択し、Rmax1とする(ステップS364)。このRmax1をESD(extream studentized deviate)と呼ぶ。CPU12は、Rmax1として選択された被験者のIDとRを逸脱データとしてメモリ20に記録する(被験者IDだけを記録してもよい)。
ここで、αは危険度であり、どの程度の棄却を行うかを決定する度合いである。危険度αは0~1の値をとり、その値が大きいほど棄却される度合いが大きくなる。したがって、危険度αが大きいほど外れ値を確実に棄却できるようになるが、反面、正常値を誤って棄却してしまうおそれも大きくなる。逆に、危険度αが小さいほど正常値を誤って棄却してしまうおそれは少なくなるが、反面、外れ値を棄却できない可能性が大きくなる。
B = SQRT((n-i-1+T2)(n-i-1))
λ = A / B
ここで、SQRT()は、()内を対象とする平方根を示す。
図10のステップS22、S23、S24に示した正中矢状断面の決定、AC-PC線の決定、AC-PC線による被験者画像と標準画像との合致処理に代えて、相互情報を用いた方法を用いてもよい。この方法においては、被験者の脳血流画像を標準画像に対して僅かずつずらし、各位置における相互情報を算出して、両画像が最もよく重なっている位置を見いだす。相互情報を用いた位置あわせについては、F. Maes et al., "Multimodality Image Registration by Maximization of Mutual Information," IEEE Transactions on Medical Imaging, (USA), 1997, 16, 2, p187-198に詳しい説明がなされている。なお、相互情報を用いた位置あわせ処理は、ワシントン大学の蓑島博士から提供されている3D-SSPプログラム中のstereo programの部分を用いて実現することができる。
元データとして、健常者のIMP投与頭部SPECTデータをもとに作成されたノーマルデータベースであるChibaDB_ver2IMP60-64AGLBMNSFM(国立大学法人 千葉大学により供給)のsurface pixel data(以下、元画像データという)を用いた。元画像データの各ピクセルにおける平均値に対し、変動係数10%の正規分布にのるランダムデータを各ピクセルごとに計算し、40セットの仮想画像データからなる仮想画像データベースを作成した(画像を、図31に示す。以下、単に仮想画像データベースという。)。この仮想画像データベースにおける各ピクセルのデータセットに対し、表1~4記載の平均値および変動係数で発生させたランダムデータを10個混入させ、検証用のデータ群を作成した(50セットの画像データに相当)。
Claims (19)
- 疾患者を含む複数人についての脳の各部位の状態を示す脳機能画像に基づいて、脳疾患を判定する際に用いる脳の各部位についての正常状態を示す代替正常脳データベースを生成する装置であって、
各人の脳機能画像の各部位における状態値を、脳機能画像の一部又は全部の状態値に基づいて正規化する手段と、
各人の脳機能画像を解剖学的標準脳に合致するように空間的な変形を行う手段と、
正規化および解剖学的標準化がなされた各人の脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、脳疾患を示すと推定される状態値を除外する手段と、
脳疾患を示すと推定される状態値を除外した脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳データベースを生成する手段とを備え、
前記除外手段は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴とする代替正常脳データベース生成装置。 - 請求項1の代替正常脳データベース生成装置において、
前記除外手段は、各部位ごとに、
i)複数人全ての状態値を対象状態値群とし、
ii)対象状態値群についてESDを算出し、逸脱状態値として記録し、
iii)ESDが臨界値λを超えているかどうかを判断し、判断結果を逸脱状態値に対応付けて記録し、
iv)繰り返しが所定回数に達していなければ、対象となっている状態値群から逸脱状態値を取り除き新しい対象状態値群を形成し、上記ii)以下を実行し、
v)所定の繰り返し回数に達すれば、最後にESDが臨界値λを超えた時の逸脱状態値より前の逸脱状態値を全て除外対象とすること
を特徴とする代替正常脳データベース生成装置。 - 請求項1または2の代替正常脳データベース生成装置において、
前記危険度αは、0.1~0.6の範囲で設定することを特徴とする代替正常脳データベース生成装置。 - 請求項3の代替正常脳データベース生成装置において、
前記危険度αは、0.2~0.5の範囲で設定することを特徴とする代替正常脳データベース生成装置。 - 請求項4の代替正常脳データベース生成装置において、
前記危険度αは、0.2~0.4の範囲で設定することを特徴とする代替正常脳データベース生成装置。 - 請求項1~5の代替正常脳データベース生成装置において、
正規化および解剖学的標準化がなされた各人の脳機能画像に基づいて、脳表から内部に垂直に所定深さまでの状態値の代表値を選択し、これを当該脳表の状態値として各人の脳表状態画像を生成する手段をさらに備え、
前記脳表状態画像において状態値が示されている部位を、前記脳疾患を判定する際に用いる脳の各部位として用いる代替正常脳データベース生成装置。 - 請求項1~6のいずれかの代替正常脳データベース生成装置において、
前記脳機能画像は、血流対応値を示す脳血流画像であることを特徴とする代替正常脳データベース生成装置。 - 請求項1~7のいずれかの代替正常脳データベース生成装置において、
前記生成手段は、脳疾患を示すと推定される状態値を除外した脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、平均状態値および標準偏差を算出し、前記正常状態を示すデータベースを生成する代替正常脳データベース生成装置。 - 疾患者を含む複数人についての脳の各部位の血流状態を示す脳血流画像に基づいて脳疾患を判定する際に用いる代替正常脳表データベースを生成する装置であって、
各人の脳血流画像の各部位における脳血流対応値を、脳血流画像の一部又は全部の脳血流対応値に基づいて正規化する手段と、
各人の脳血流画像を解剖学的標準脳に合致するように空間的な変形を行う手段と、
正規化および解剖学的標準化がなされた各人の脳血流画像に基づいて、脳表から内部に垂直に所定深さまでの脳血流対応値の代表値を選択し、これを当該脳表の血流対応値として各人の脳表血流画像を生成する手段と、
脳表血流画像生成手段によって生成された各人の脳表血流画像の各部位ごとに、脳疾患を示すと推定される血流対応値を除外する手段と、
脳疾患を示すと推定される血流対応値を除外した脳表血流画像に基づいて、前記脳疾患を判定する際に用いる脳表の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳表データベースを生成する手段と、
を備えた代替正常脳表データベース生成装置であって、
前記除外手段は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴とする代替正常脳表データベース生成装置。 - 疾患者を含む複数人についての脳の各部位の状態を示す脳機能画像に基づいて、脳疾患を判定する際に用いる脳の各部位についての正常状態を示す代替正常脳データベースを生成する装置をコンピュータによって実現するためのプログラムであって、前記プログラムは、
各人の脳機能画像の各部位における状態値を、脳機能画像の一部又は全部の状態値に基づいて正規化する手段と、
各人の脳機能画像を解剖学的標準脳に合致するように空間的な変形を行う手段と、
正規化および解剖学的標準化がなされた各人の脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、脳疾患を示すと推定される状態値を除外する手段と、
脳疾患を示すと推定される状態値を除外した脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳データベースを生成する手段とをコンピュータによって実現し、
前記除外手段は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴とする代替正常脳データベース生成プログラム。 - 請求項10の代替正常脳データベース生成プログラムにおいて、
前記除外手段は、各部位ごとに、
i)複数人全ての状態値を対象状態値群とし、
ii)対象状態値群についてESDを算出し、逸脱状態値として記録し、
iii)ESDが臨界値λを超えているかどうかを判断し、判断結果を逸脱状態値に対応付けて記録し、
iv)繰り返しが所定回数に達していなければ、対象となっている状態値群から逸脱状態値を取り除き新しい対象状態値群を形成し、上記ii)以下を実行し、
v)所定の繰り返し回数に達すれば、最後にESDが臨界値λを超えた時の逸脱状態値より前の逸脱状態値を全て除外対象とすること
を特徴とする代替正常脳データベース生成プログラム。 - 請求項10または11の代替正常脳データベース生成プログラムにおいて、
前記危険度αは、0.1~0.6の範囲で設定することを特徴とする代替正常脳データベース生成プログラム。 - 請求項12の代替正常脳データベース生成プログラムにおいて、
前記危険度αは、0.2~0.5の範囲で設定することを特徴とする代替正常脳データベース生成プログラム。 - 請求項13の代替正常脳データベース生成プログラムにおいて、
前記危険度αは、0.2~0.4の範囲で設定することを特徴とする代替正常脳データベース生成プログラム。 - 請求項10~14の代替正常脳データベース生成プログラムにおいて、
正規化および解剖学的標準化がなされた各人の脳機能画像に基づいて、脳表から内部に垂直に所定深さまでの状態値の代表値を選択し、これを当該脳表の状態値として各人の脳表状態画像を生成する手段をさらに備え、
前記脳表状態画像において状態値が示されている部位を、前記脳疾患を判定する際に用いる脳の各部位として用いる代替正常脳データベース生成プログラム。 - 請求項10~15のいずれかの代替正常脳データベース生成プログラムにおいて、
前記脳機能画像は、血流対応値を示す脳血流画像であることを特徴とする代替正常脳データベース生成プログラム。 - 請求項10~16のいずれかの代替正常脳データベース生成プログラムにおいて、
前記生成手段は、脳疾患を示すと推定される状態値を除外した脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、平均状態値および標準偏差を算出し、前記正常状態を示すデータベースを生成する代替正常脳データベース生成プログラム。 - 疾患者を含む複数人についての脳の各部位の血流状態を示す脳血流画像に基づいて脳疾患を判定する際に用いる代替正常脳表データベースを生成する装置をコンピュータによって実現するためのプログラムであって、前記プログラムは、
各人の脳血流画像の各部位における脳血流対応値を、脳血流画像の一部又は全部の脳血流対応値に基づいて正規化する手段と、
各人の脳血流画像を解剖学的標準脳に合致するように空間的な変形を行う手段と、
正規化および解剖学的標準化がなされた各人の脳血流画像に基づいて、脳表から内部に垂直に所定深さまでの脳血流対応値の代表値を選択し、これを当該脳表の血流対応値として各人の脳表血流画像を生成する手段と、
脳表血流画像生成手段によって生成された各人の脳表血流画像の各部位ごとに、脳疾患を示すと推定される血流対応値を除外する手段と、
脳疾患を示すと推定される血流対応値を除外した脳表血流画像に基づいて、前記脳疾患を判定する際に用いる脳表の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳データベースを生成する手段と、
をコンピュータによって実現するものであり、
前記除外手段は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴とする代替正常脳表データベース生成プログラム。 - 疾患者を含む複数人についての脳の各部位の状態を示す脳機能画像に基づいて、脳疾患を判定する際に用いる脳の各部位についての正常状態を示す代替正常脳データベースを生成する方法であって、
各人の脳機能画像の各部位における状態値を、脳機能画像の一部又は全部の状態値に基づいて正規化し、
各人の脳機能画像を解剖学的標準脳に合致するように空間的な変形を行い、
正規化および解剖学的標準化がなされた各人の脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、脳疾患を示すと推定される状態値を除外し、
脳疾患を示すと推定される状態値を除外した脳機能画像に基づいて、前記脳疾患を判定する際に用いる脳の各部位ごとに、少なくとも平均状態値を算出し、前記正常状態を示す代替正常脳データベースを生成する方法において、
前記除外処理は、部位ごとに、複数人の状態値について偏差/標準偏差のうちの最大値をESD(extream studentized deviate)として算出し、当該ESDが所定の危険度αに基づいて算出した臨界値λを超えている場合には、当該ESDとして選択された状態値を、少なくとも除外することを特徴とする代替正常脳データベース生成方法。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2009294085A AU2009294085B2 (en) | 2008-09-22 | 2009-09-17 | Device for creating database of alternative normal brain |
JP2010529646A JP5404633B2 (ja) | 2008-09-22 | 2009-09-17 | 代替正常脳データベース生成装置、代替正常脳データベース生成プログラムおよび代替正常脳データベース生成方法 |
CA2738124A CA2738124A1 (en) | 2008-09-22 | 2009-09-17 | Device for creating database of alternative normal brain |
EP09814317A EP2347711A4 (en) | 2008-09-22 | 2009-09-17 | DEVICE FOR GENERATING A DATABASE OF AN ALTERNATIVE NORMAL BRAIN |
US13/119,914 US8437523B2 (en) | 2008-09-22 | 2009-09-17 | Device for creating database of alternative normal brain |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2008-242196 | 2008-09-22 | ||
JP2008242196 | 2008-09-22 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010032472A1 true WO2010032472A1 (ja) | 2010-03-25 |
Family
ID=42039319
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/004693 WO2010032472A1 (ja) | 2008-09-22 | 2009-09-17 | 代替正常脳データベース生成装置 |
Country Status (6)
Country | Link |
---|---|
US (1) | US8437523B2 (ja) |
EP (1) | EP2347711A4 (ja) |
JP (1) | JP5404633B2 (ja) |
AU (1) | AU2009294085B2 (ja) |
CA (1) | CA2738124A1 (ja) |
WO (1) | WO2010032472A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012063939A1 (ja) * | 2010-11-12 | 2012-05-18 | 株式会社東芝 | 画像診断装置 |
JP2015034779A (ja) * | 2013-08-09 | 2015-02-19 | 株式会社東芝 | 核医学診断装置 |
JPWO2014034724A1 (ja) * | 2012-08-30 | 2016-08-08 | 株式会社島津製作所 | 脳断層動態画像解析装置 |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012173976A2 (en) * | 2011-06-17 | 2012-12-20 | Carroll Robert G | Methods and apparatus for assessing activity of an organ and uses thereof |
CN104093354B (zh) * | 2011-12-15 | 2017-06-09 | 联邦科学与工业研究组织 | 用于评估医学图像的方法和设备 |
US9332954B2 (en) * | 2013-03-15 | 2016-05-10 | Jacqueline K. Vestevich | Systems and methods for evaluating a brain scan |
CN103425999A (zh) * | 2013-08-27 | 2013-12-04 | 西安电子科技大学 | 基于非负张量投影算子分解算法的脑认知状态判定方法 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA2469480A1 (en) * | 2001-12-27 | 2003-07-17 | Agy Therapeutics, Inc. | Use of biomolecular targets in the treatment and visualization of tumors |
JP2004239782A (ja) * | 2003-02-06 | 2004-08-26 | Daiichi Radioisotope Labs Ltd | 脳画像データ処理システム、方法、プログラムおよび記録媒体 |
JP4335736B2 (ja) * | 2003-04-23 | 2009-09-30 | 富士フイルムRiファーマ株式会社 | 脳画像データ処理システム、方法、プログラムおよび記録媒体 |
JP4581088B2 (ja) * | 2005-05-17 | 2010-11-17 | 国立大学法人 筑波大学 | 計算機支援診断装置および方法 |
US20100074480A1 (en) | 2008-09-22 | 2010-03-25 | University Of Washington | Device for generating alternative of normal brain database |
-
2009
- 2009-09-17 CA CA2738124A patent/CA2738124A1/en not_active Abandoned
- 2009-09-17 JP JP2010529646A patent/JP5404633B2/ja active Active
- 2009-09-17 AU AU2009294085A patent/AU2009294085B2/en not_active Ceased
- 2009-09-17 WO PCT/JP2009/004693 patent/WO2010032472A1/ja active Application Filing
- 2009-09-17 US US13/119,914 patent/US8437523B2/en not_active Expired - Fee Related
- 2009-09-17 EP EP09814317A patent/EP2347711A4/en not_active Withdrawn
Non-Patent Citations (3)
Title |
---|
See also references of EP2347711A4 * |
TAICHI MATSUO ET AL.: "253 No Ketsuryu SPECT Tokeigakuteki Kaisekiho ni Okeru Seijo Database Sakusei ni Tsuite no Kento", DAI 35 KAI JAPANESE SOCIETY OF RADIOLOGICAL TECHNOLOGY SHUKI GAKUJUTSU TAIKAI IPPAN KENKYU HAPPYO YOKOSHU, vol. 63, no. 9, 20 September 2007 (2007-09-20), pages 1049, XP008138422 * |
YASUSHI YAMAMOTO ET AL.: "No Ketsuryu SPECT Tokeigakuteki Gazo Kaiseki ni Okeru Seido - Hyoka ni Kansuru Kenkyuhan Hokoku", JAPANESE JOURNAL OF RADIOLOGICAL TECHNOLOGY, vol. 64, no. 6, 20 June 2008 (2008-06-20), pages 752 - 765, XP008137096 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012063939A1 (ja) * | 2010-11-12 | 2012-05-18 | 株式会社東芝 | 画像診断装置 |
JP2012115658A (ja) * | 2010-11-12 | 2012-06-21 | Toshiba Corp | 画像診断装置 |
CN102630151A (zh) * | 2010-11-12 | 2012-08-08 | 株式会社东芝 | 图像诊断装置 |
US8903146B2 (en) | 2010-11-12 | 2014-12-02 | Kabushiki Kaisha Toshiba | Image diagnosis apparatus |
JP2018011983A (ja) * | 2010-11-12 | 2018-01-25 | 東芝メディカルシステムズ株式会社 | 画像診断装置 |
JPWO2014034724A1 (ja) * | 2012-08-30 | 2016-08-08 | 株式会社島津製作所 | 脳断層動態画像解析装置 |
US9715728B2 (en) | 2012-08-30 | 2017-07-25 | Shimadzu Corporation | Brain tomographic dynamic image analyzing apparatus |
JP2015034779A (ja) * | 2013-08-09 | 2015-02-19 | 株式会社東芝 | 核医学診断装置 |
Also Published As
Publication number | Publication date |
---|---|
EP2347711A4 (en) | 2013-03-13 |
AU2009294085A2 (en) | 2011-07-28 |
JPWO2010032472A1 (ja) | 2012-02-09 |
CA2738124A1 (en) | 2010-03-25 |
JP5404633B2 (ja) | 2014-02-05 |
AU2009294085B2 (en) | 2013-08-22 |
US8437523B2 (en) | 2013-05-07 |
US20110172522A1 (en) | 2011-07-14 |
AU2009294085A1 (en) | 2010-03-25 |
EP2347711A1 (en) | 2011-07-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5367066B2 (ja) | 代替正常脳データベース生成装置 | |
JP5424902B2 (ja) | Pet/mrフロー推定を用いて補われる自動診断及び自動整列 | |
JP5404633B2 (ja) | 代替正常脳データベース生成装置、代替正常脳データベース生成プログラムおよび代替正常脳データベース生成方法 | |
JP5243865B2 (ja) | 脳疾患診断システム | |
ES2414614T3 (es) | Herramientas para ayudar en el diagnóstico de enfermedades neurodegenerativas | |
US8170347B2 (en) | ROI-based assessment of abnormality using transformation invariant features | |
US10888290B2 (en) | Medical-image processing apparatus and medical-image diagnostic apparatus | |
US8311359B2 (en) | Registering brain images by aligning reference ellipses | |
WO2006025963A2 (en) | Method, system and storage medium which includes instruction for analyzing anatomical structures | |
US10910101B2 (en) | Image diagnosis support apparatus, image diagnosis support method, and image diagnosis support program | |
Sun et al. | Automated template-based PET region of interest analyses in the aging brain | |
US20080273775A1 (en) | Cartesian human morpho-informatic system | |
Kudo et al. | Image processing method for analyzing cerebral blood-flow using SPECT and MRI | |
Qian et al. | A fast and automatic approach to extract the brain and midsagittal lines from FDG-PET head scans | |
JP2018044849A (ja) | コントロールデータベース生成装置、コントロールデータベース生成方法、およびプログラム | |
CA2694527A1 (en) | Cartesian human morpho-informatic system | |
EP1784784A2 (en) | Method, system and storage medium which includes instruction for analyzing anatomical structures |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09814317 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010529646 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2009294085 Country of ref document: AU |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13119914 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2738124 Country of ref document: CA |
|
ENP | Entry into the national phase |
Ref document number: 2009294085 Country of ref document: AU Date of ref document: 20090917 Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2009814317 Country of ref document: EP |