JP4721693B2 - Intracranial volume and local brain structure analysis program, recording medium, and intracranial volume and local brain structure analysis method - Google Patents

Description

The present invention relates to an intracranial volume and a local brain structure analysis program for causing a computer to process brain image data, and more particularly, based on brain image data of a subject's brain measured by computer tomography. Intracranial volume and local brain structure analysis program for automatically setting a region of interest for each internal or anatomical part of the brain and causing a computer to execute processing for automatically measuring the volume of the region of interest Etc.

  In recent years, diagnostic imaging has advanced remarkably with the advance of medicine. Computer imaging techniques for capturing the state of the body as an image and making an accurate diagnosis without imposing a heavy burden on the patient, such as magnetic resonance imaging or magnetic resonance imaging (MRI), X-ray tomography or X-ray tomography (CT), ultrasonography or ultrasonography, and radiography or radiography are essential in current medical settings. (See Non-Patent Document 1). Image diagnosis using the above-described computer image diagnosis method is widely used as providing functional information such as early diagnosis of a disease of a patient, selection of a treatment method, prediction and determination of a treatment effect, and the like. In particular, since an image by MRI (MRI image) and an image by CT (CT image) are anatomical images representing the morphology of the brain, they are generally images with high resolution and clear positional information in the brain. is there.

  Examining the human brain volume in a living state is important in neuroscience, medicine, and medicine. By using the above-described MRI, the form of the human brain can be depicted as an image in a living state. In the medical field, this MRI image is used to distinguish and diagnose a brain that has atrophy due to a disease such as Alzheimer's disease (AD; see Non-Patent Document 2) and a normal person's brain. .

“What is diagnostic imaging?” [Online], Daiichi Radioisotope Research Institute, Inc. [searched on November 8, 2004], Internet <URL: http://www.drl.co.jp/html/ naruhodo.htm> Co-authored by Eizo Iseki and Kenji Kosaka, “Clinical Psychiatry S9 Alzheimer's Disease”, pp. 31-41, “Neuropathology A. Characteristics of brain lesions”, 2000, published by Nakayama Shoten.

  As described above, when performing brain discrimination and diagnosis using MRI images, it is necessary to measure the volume of the brain in order to quantitatively evaluate the degree of brain atrophy. Specifically, a region of interest is set in the brain region, and the volume of the portion where the region of interest is set is measured. Conventionally, the setting of the region of interest in the brain region and the measurement of the volume of the brain of the portion where the region of interest is set depend on manual operation, and experience and a great deal of time have been required. Conventionally, a method for automatically setting a region of interest and measuring a volume at the level of the whole brain, cerebrum, or cerebellum has been devised. However, traditional methods are not only inaccurate, but also set the region of interest and measure the volume in a finer local area of the brain, such as the temporal lobe, or finer hippocampus or hippocampal wrapping Regions of interest and volume measurements for brain structures such as (parahippocampal grrus) could not be performed. In recent years, methods for semi-automatically setting the region of interest and measuring the volume of the hippocampus through some manual operation have been devised, but these operations are still complicated and have been put to practical use in clinical settings. There was no problem.

Accordingly, an object of the present invention has been made to solve the above-mentioned problems, and not only at the level of the whole brain, the cerebrum, or the cerebellum but also automatically and accurately for each finer anatomical part of the brain. An object of the present invention is to provide an intracranial volume and local brain structure analysis program that can set a region of interest and can be used practically in a clinical field.

The second object of the present invention is to provide an intracranial volume that can automatically and accurately measure the volume of the region of interest set as described above, and that can be used practically in a clinical field. The purpose is to provide a local brain structure analysis program and the like.

The intracranial volume and local brain structure analysis program of the present invention is an intracranial volume and local brain structure analysis program for causing a computer to process brain image data, and is measured by a computer tomography method on a computer. An extraction step of extracting brain image data of the subject's brain within a predetermined range from the brain image data of the subject's brain recorded in the brain image data recording unit, and brain image data of the subject's brain extracted in the extraction step An image superposition step for superimposing an anterior commissure-rear commissure line in the vicinity of the anterior commissure-rear commissure line in the brain image data of the standard brain; If the intracranial volume is not calculated from the brain image data of the brain, at least the gray matter brain image Separation step of recording data on the brain image data recording unit after separation separates the region of interest setting step of recording the region of interest recording unit sets a region of interest in the brain image data of the standard brain, in the region of interest recording unit The recorded brain image data of the standard brain is deformed based on the brain image data of the gray matter portion in the standard brain and the brain image data of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit. By performing reverse anatomical normalization to match the brain image data of the standard brain in which the region of interest is set with the brain image data of the subject, the region of interest is converted into the brain image data of the subject's brain. An intracranial volume and local brain structure analysis program for executing a region-of-interest conversion step recorded in the region-of-interest recording unit.

Here, in the intracranial volume and local brain structure analysis program of the present invention , prior to the extraction step, the brain of the subject measured by the computer tomography method and recorded in the first brain image data recording unit in the first format Further comprising a format conversion step of converting the brain image data into a second format and recording it in a second brain image data recording unit , wherein the extraction step is recorded in the second brain image data recording unit in the format conversion step. It was identified parietal brain image data of a subject's brain, brain image data of the subject's brain in a predetermined range from the identified said head top can be extracted.

Here, in the intracranial volume and local brain structure analysis program of the present invention, the brain image data of the gray matter portion in the brain of the subject recorded in the post-separation brain image data recording unit prior to the region of interest conversion step Can be further provided with a smoothing step of applying predetermined smoothing to the post-separation brain image data recording unit.

Here, in the intracranial volume and local brain structure analysis program according to the present invention, the computed tomography in the preprocessing step is magnetic resonance imaging, and is recorded in the region of interest recording unit in the region of interest conversion step. A volume calculating step of calculating a volume of the region of interest based on the region of interest converted into the brain image data of the brain of the subject can be further provided.

Here, in the intracranial volume and local brain structure analysis program according to the present invention, the brain image data acquired in the preprocessing step and recorded in the post-separation brain image data recording unit are the separated gray matter portions. Brain image data, white matter brain image data, and cerebrospinal fluid brain image data, and the region of interest setting step sets an intracranial region of interest within the skull in the standard brain brain image data. The brain image data of the standard brain is recorded in the region-of-interest recording unit, and the region-of-interest conversion step includes the brain image data of the standard brain in which the intracranial region of interest recorded in the region-of-interest recording unit is set in the standard brain. By transforming the brain image data of the gray matter portion and the brain image data of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit, the subject's brain The region of interest within the skull is converted into brain image data and recorded in the region of interest recording unit, and the volume calculating step uses the brain image data recorded in the post-separation brain image data recording unit, Intracranial identification brain image data for identifying the intracranial region of interest is created from the brain image data of the subject's brain converted from the intracranial region of interest recorded in the region of interest recording unit, and is recorded in the identification brain image data recording unit. Intracranial identification brain image data creation step, gray matter brain image data and white matter brain image data recorded in the post-separation brain image data recording unit, and skull recorded in the identification brain image data recording unit Based on the internal identification brain image data, all gray matter and all white matter identification brain image data for identifying all gray matter and all white matter are created and recorded in the identification brain image data recording unit. Based on the brain image data creation step, the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, and the brain image data of the white matter portion recorded in the separated brain image data recording unit In the case of calculating the whole gray matter identification brain image data for creating the whole gray matter identification brain image data for identifying the whole gray matter and recording it in the identification brain image data recording unit, and calculating the volume of the intracranial region of interest, Calculating the volume of the intracranial region of interest by multiplying the total voxel value in the intracranial region of interest identified by the intracranial region of identification brain image data recorded in the discriminating brain image data recording unit by a predetermined numerical value; and Whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, gray matter brain image data and white matter brain image recorded in the separated brain image data recording unit Based on the image data, find the total voxel value in the whole gray matter and the whole white matter region, and calculate the volume of the whole gray matter and the whole white matter region by multiplying the total voxel value by a predetermined numerical value. -Total white matter region volume calculation step, and in the total gray matter region identified by the whole gray matter identification brain image data recorded in the identification brain image data recording unit, recorded in the post-separation brain image data recording unit A total gray matter region volume calculating step of calculating a total value of voxel values in the voxels corresponding to the gray matter brain image data and calculating the volume of the total gray matter region by multiplying the total value by a predetermined numerical value; The total gray matter region volume calculating step determines the total gray matter region volume calculation step from the total gray matter and total white matter region total voxel values determined in the total gray matter region A total white matter region volume calculating step of calculating a total voxel value in the total white matter region by subtracting the measured value and calculating a volume of the total white matter region by multiplying the total voxel value by a predetermined numerical value. Can do.

Here, in the intracranial volume and local brain structure analysis program according to the present invention, the total gray matter / total white matter region volume calculating step includes the whole gray matter / total white matter discrimination brain recorded in the discrimination brain image data recording unit. In the contour portion of the image data, the larger voxel value of the voxel values of the voxels corresponding to the brain image data of the gray matter portion and the brain image data of the white matter portion recorded in the post-separation brain image data recording portion is used. By using the voxel value 1 in the whole gray matter / whole white matter identification brain image data, the total voxel value in the whole gray matter and the whole white matter region is obtained, and the total voxel value is multiplied by a predetermined numerical value. The total gray matter and the volume of the whole white matter region can be calculated.

Here, in the intracranial volume and local brain structure analysis program of the present invention, the region-of-interest setting step sets an intracranial region of interest in the entire skull in the brain image data of the standard brain, and another standard brain A part of the region of interest within the skull is set in the brain image data of the brain, and the brain image data of each standard brain is recorded in the region of interest recording unit, and the region of interest converting step is performed in the region of interest recording unit The recorded brain image data of the standard brain in which the region of interest within the skull is set are the brain image data of the gray matter portion in the standard brain and the gray matter portion in the brain of the subject recorded in the post-separation brain image data recording unit. By transforming based on the brain image data, the intracranial region of interest is converted into the brain image data of the subject's brain and recorded in the region of interest recording unit, and recorded in the region of interest recording unit Brain image data of a standard brain in which a local region of interest is set, brain image data of a gray matter portion in the standard brain, and brain image data of a gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit The local region of interest is converted into the brain image data of the subject's brain and is recorded in the region of interest recording unit, and the volume calculating step calculates the volume of the local region of interest , All gray matter and all white matter regions identified by the whole gray matter / white matter discrimination brain image data created in the all gray matter / whole white matter discrimination brain image data creation step and recorded in the discrimination brain image data recording unit In the common area with the local region of interest indicated by the brain image data of the subject's brain converted from the local region of interest recorded in the region of interest recording unit in the region of interest conversion step, Based on the brain image data of the gray matter portion and the brain image data of the white matter portion recorded in the post-separation brain image data recording unit, a total voxel value of gray matter and white matter in the local region of interest is obtained, and the total voxel value The gray matter and white matter volume calculating step for calculating the gray matter and white matter volume in the local region of interest by multiplying by the predetermined numerical value and the whole gray matter discriminating brain image data creating step and the discriminating brain image data The brain of the subject's brain in which the region of the whole gray matter identified by the whole gray matter identification brain image data recorded in the recording unit and the local region of interest recorded in the region of interest recording unit in the region of interest conversion step are converted A box corresponding to the brain image data of the gray matter portion recorded in the post-separation brain image data recording unit in the common area with the local region of interest indicated by the image data. A local gray matter volume calculating step for calculating a gray matter volume in the local region of interest by multiplying the total value by a predetermined numerical value; and the local gray matter / white matter volume calculating step The total voxel value in the white matter region of the local region of interest is obtained by subtracting the total value obtained in the local gray matter volume calculating step from the total gray matter and white matter voxel values in the local region of interest obtained in step (b). And a local white matter region volume calculating step of calculating a white matter volume in the local region of interest by multiplying the total voxel value by a predetermined numerical value.

Here, in the intracranial volume and local brain structure analysis program according to the present invention, the local gray matter / white matter volume calculating step includes the whole gray matter / white matter discrimination brain image data recorded in the discrimination brain image data recording unit. In the common region of the whole gray matter and the whole white matter region identified by the region of interest and the local region of interest indicated by the brain image data of the subject's brain converted from the region of interest recorded in the region of interest recording unit. In the contour portion of the region, the larger voxel value of the voxel values of the corresponding voxels of the gray matter portion brain image data and the white matter portion brain image data recorded in the post-separation brain image data recording portion is used, By using the voxel value 1 inside the common region, the total voxels in the gray matter and white matter regions in the local region of interest are obtained, and the total voxel value is set to a predetermined value. It is possible to calculate the gray matter and white matter of the volume in said local region of interest by multiplying the value.

Here, in the intracranial volume and local brain structure analysis program according to the present invention, the intracranial identification brain image data creating step includes the gray matter brain image data and white matter recorded in the post-separation brain image data recording unit. A brain image data adding step for adding the brain image data of the portion and the brain image data of the cerebrospinal fluid portion, and brain image data smoothing for performing predetermined smoothing on the brain image data added in the brain image data adding step A brain image data binarization step for binarizing the brain image data smoothed in the brain image data smoothing step based on a predetermined threshold; By calculating the logical product of the brain image data that has been converted into the brain image data of the brain of the subject whose intracranial region of interest has been converted in the region of interest conversion step, the skull is obtained. You can create intracranial identification brain image data for identifying the region of interest and a intracranial logical step of recording the identification brain image data recording unit.

Here, in the intracranial volume and local brain structure analysis program of the present invention, the whole gray matter / whole white matter identification brain image data creation step includes the step of producing the brain of the gray matter portion recorded in the post-separation brain image data recording unit. The gray matter partial brain image data binarization step for binarizing the image data based on a predetermined threshold, and the white matter brain image data recorded in the post-separation brain image data recording unit 2 based on the predetermined threshold White matter partial brain image data binarization step for performing binarization, gray matter brain image data binarized by the gray matter partial brain image data binarization step, and white matter partial brain image data binarization step A gray matter / white matter partial brain image data logical sum step for obtaining a logical sum with the brain image data of the white matter portion binarized in step (b), and the gray matter / white matter partial brain image data logical sum step. In the brain image data for which logical sum is obtained in the step, an embedding step for embedding a predetermined portion with a single value, a brain image data subjected to an embedding operation in the embedding step, and the identification brain image data recording unit By taking the logical product with the recorded intracranial identification brain image data, all gray matter and all white matter identification brain image data is created and recorded in the identification brain image data recording unit. And the whole gray matter discrimination brain image data creating step includes the white matter portion brain image data binarized in the white matter partial brain image data binarization step. From the white matter partial embedding step for performing an operation of embedding a predetermined portion with one value, and from the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, A whole gray matter difference step of creating a whole gray matter identification brain image data and recording it in the identification brain image data recording unit by subtracting the brain image data of the white matter portion subjected to the embedding operation in the matter portion embedding step; Can be provided.

The recording medium of the present invention is a computer-readable recording medium on which any one or a plurality of intracranial volumes and a local brain structure analysis program of the present invention are recorded.

The intracranial volume and local brain structure analysis method according to the present invention is an intracranial volume and local brain structure analysis method for causing a computer to perform processing of brain image data, and is measured by a computer tomography method. An extraction step for extracting brain image data of the subject's brain within a predetermined range for the brain image data of the subject's brain recorded in the data recording unit, and a previous step in the brain image data of the subject's brain extracted in the extraction step An image superimposing step for superimposing the commissure-posterior commissural line to the vicinity of the anterior commissural-posterior commissural line in the brain image data of the standard brain, and the brain of the subject on which image superposition is performed in the image superimposing step If the intracranial volume is not calculated from the brain image data of the brain, at least the brain image data of the gray matter portion is separated and separated A separation step of recording the image data recording unit, and the region of interest setting step of recording the region of interest recording unit sets a region of interest in the brain image data of the standard brain, the standard brain recorded in the region of interest recording unit By transforming the brain image data based on the brain image data of the gray matter portion in the standard brain and the brain image data of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit, the region of interest is Reverse anatomical normalization that matches the set brain image data of the standard brain with the brain image data of the subject, converts the region of interest into the brain image data of the subject's brain , and the region of interest recording unit And a region-of-interest conversion step for recording the information.

Here, in the intracranial volume and local brain structure analysis method of the present invention , prior to the extraction step, the brain of the subject measured by the computer tomography method and recorded in the first brain image data recording unit in the first format Further comprising a format conversion step of converting the brain image data into a second format and recording it in a second brain image data recording unit , wherein the extraction step is recorded in the second brain image data recording unit in the format conversion step. It was identified parietal brain image data of a subject's brain, brain image data of the subject's brain in a predetermined range from the identified said head top can be extracted.

Here, in the intracranial volume and local brain structure analysis method of the present invention, the brain image data of the gray matter portion in the brain of the subject recorded in the post-separation brain image data recording unit prior to the region of interest conversion step And a smoothing step of performing a predetermined smoothing on the post-separation brain image data recording unit.

Here, in the intracranial volume and local brain structure analysis method of the present invention, the computed tomography in the preprocessing step is magnetic resonance imaging, and is recorded in the region of interest recording unit in the region of interest conversion step. A volume calculating step of calculating a volume of the region of interest based on the region of interest converted into the brain image data of the brain of the subject can be further provided.

Here, in the intracranial volume and local brain structure analysis method according to the present invention, the brain image data acquired in the preprocessing step and recorded in the post-separation brain image data recording unit are the gray matter portions that are separated. Including brain image data, brain image data of a white matter portion, and brain image data of a cerebrospinal fluid portion, and the region-of-interest setting step sets an intracranial region of interest within the skull in the brain image data of a standard brain , Setting a part of the local region of interest in the skull in the brain image data of another standard brain, recording the brain image data of each standard brain in the region of interest recording unit, the region of interest conversion step, The brain image data of the standard brain in which the intracranial region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the brain of the subject recorded in the post-separation brain image data recording unit In And transforming the region of interest within the skull into the brain image data of the subject's brain, and recording the region of interest in the region of interest recording unit. The brain image data of the standard brain in which the local region of interest recorded is set, the brain image data of the gray matter portion in the standard brain, and the gray matter portion brain in the brain of the subject recorded in the post-separation brain image data recording unit The local region of interest is transformed into the brain image data of the subject's brain by being deformed based on the image data and recorded in the region of interest recording unit, and the volume calculating step includes the post-separation brain image data recording By using the brain image data recorded in the region, the intracranial information for identifying the intracranial region of interest from the brain image data of the subject's brain converted from the intracranial region of interest recorded in the region of interest recording unit. Intracranial identification brain image data creation step for creating brain image data and recording it in the identification brain image data recording unit, gray matter brain image data and white matter brain recorded in the post-separation brain image data recording unit Based on the image data and the intracranial identification brain image data recorded in the identification brain image data recording unit, the whole gray matter and the whole white matter identification brain image data for identifying the whole gray matter and the whole white matter are created and the identification is performed. Whole gray matter / whole white matter discrimination brain image data creation step to be recorded in the brain image data recording unit, whole gray matter / whole white matter discrimination brain image data and the separated brain image data recorded in the discrimination brain image data recording unit Based on the brain image data of the white matter portion recorded in the recording unit, the whole gray matter identifying brain image data for identifying the whole gray matter is generated and recorded in the identifying brain image data recording unit. And when calculating the volume of the intracranial region of interest, a predetermined numerical value is set as the total voxel value in the intracranial region of interest identified by the intracranial discernment brain image data recorded in the discriminating brain image data recording unit. To calculate the volume of the region of interest within the skull, and the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit and the post-separation brain image data recording unit Based on the brain image data of the gray matter part and the brain image data of the white matter part, a total voxel value in the whole gray matter and the whole white matter region is obtained, and the total gray matter and Whole gray matter / whole white matter region volume calculating step for calculating the volume of the whole white matter region, and the whole gray matter identified by the whole gray matter identification brain image data recorded in the discrimination brain image data recording unit In the region, a total value of voxel values in voxels corresponding to the brain image data of the gray matter portion recorded in the post-separation brain image data recording unit is obtained, and the total gray value is obtained by multiplying the total value by a predetermined value. The total gray matter region from the total gray matter and total white matter regions obtained in the total gray matter / total white matter region volume calculation step and the total gray matter region volume calculation step for calculating the volume of the quality region The total white matter that calculates the total voxel value in the total white matter region by subtracting the total value obtained in the volume calculating step and calculates the volume of the total white matter region by multiplying the total voxel value by a predetermined value A region volume calculating step.

Here, in the intracranial volume and local brain structure analysis method according to the present invention, the total gray matter / total white matter region volume calculating step includes the whole gray matter / total white matter discrimination brain recorded in the discrimination brain image data recording unit. In the contour portion of the image data, the larger voxel value of the voxel values of the voxels corresponding to the brain image data of the gray matter portion and the brain image data of the white matter portion recorded in the post-separation brain image data recording portion is used. The total gray matter / white matter discrimination brain image data uses a voxel value of 1 to obtain a total voxel in the whole gray matter and the whole white matter region and multiply the total voxel value by a predetermined numerical value to obtain the total voxel value. The volume of the whole gray matter and the whole white matter region can be calculated.

Here, in the intracranial volume and local brain structure analysis method of the present invention, when calculating the volume of the region of local interest, the discriminated brain image data is created in the whole gray matter / whole white matter discriminating brain image data creating step. The whole gray matter and the whole white matter region identified by the whole gray matter / whole white matter identification brain image data recorded in the recording unit and the local region of interest recorded in the region of interest recording unit in the region of interest conversion step are converted. Based on the brain image data of the gray matter portion and the brain image data of the white matter portion recorded in the post-separation brain image data recording unit in a common area with the local region of interest indicated by the brain image data of the subject's brain, A total voxel value of gray matter and white matter in the local region of interest is obtained, and the amount of gray matter and white matter in the local region of interest is obtained by multiplying the total voxel value by a predetermined numerical value. And the whole gray matter identified by the whole gray matter identification brain image data created in the whole gray matter discrimination brain image data creation step and recorded in the discrimination brain image data recording unit. A post-separated brain image in a common region of a region of interest and a local region of interest indicated by brain image data of the subject's brain converted from the region of interest recorded in the region of interest recording unit in the region of interest conversion step The sum of the voxel values in the voxels corresponding to the brain image data of the gray matter portion recorded in the data recording unit is calculated, and the volume of the gray matter in the local region of interest is calculated by multiplying the total value by a predetermined value Total gray matter and white matter voxels in the local region of interest determined in the local gray matter volume calculating step and the local gray matter / white matter volume calculating step From the above, the total value obtained in the local gray matter volume calculating step is subtracted to obtain a total voxel value in the white matter region of the local region of interest, and the total interest value is multiplied by a predetermined numerical value. And a local white matter region volume calculating step of calculating a white matter volume in the region.

Here, in the intracranial volume and local brain structure analysis method according to the present invention, the local gray matter / white matter volume calculating step includes the whole gray matter / white matter discrimination brain image data recorded in the discrimination brain image data recording unit. In the common region of the whole gray matter and the whole white matter region identified by the region of interest and the local region of interest indicated by the brain image data of the subject's brain converted from the region of interest recorded in the region of interest recording unit. In the contour portion of the region, the larger voxel value of the voxel values of the corresponding voxels of the gray matter portion brain image data and the white matter portion brain image data recorded in the post-separation brain image data recording portion is used, By using a voxel value of 1 inside the common region, a total voxel value in the gray matter and white matter regions in the local region of interest is obtained, and a predetermined numerical value is obtained as the total voxel value. Gray matter and white matter of the volume in said local region of interest by multiplying can be calculated.

Here, in the intracranial volume and local brain structure analysis method of the present invention, the intracranial identification brain image data creation step includes the gray matter portion brain image data and white matter recorded in the post-separation brain image data recording unit. A brain image data adding step for adding the brain image data of the portion and the brain image data of the cerebrospinal fluid portion, and brain image data smoothing for performing predetermined smoothing on the brain image data added in the brain image data adding step A brain image data binarization step for binarizing the brain image data smoothed in the brain image data smoothing step based on a predetermined threshold; Intracranial interest is obtained by taking a logical product of the digitized brain image data and the brain image data of the subject's brain whose intracranial region of interest has been converted in the region of interest conversion step. You can create intracranial identification brain image data for identifying the frequency and a intracranial logical step of recording the identification brain image data recording unit.

Here, in the intracranial volume and local brain structure analysis method of the present invention, the whole gray matter / whole white matter discriminating brain image data creating step includes the gray matter portion of the brain recorded in the post-separation brain image data recording unit. The gray matter partial brain image data binarization step for binarizing the image data based on a predetermined threshold, and the white matter brain image data recorded in the post-separation brain image data recording unit 2 based on the predetermined threshold White matter partial brain image data binarization step for performing binarization, gray matter brain image data binarized by the gray matter partial brain image data binarization step, and white matter partial brain image data binarization step The gray matter / white matter partial brain image data logical sum step for obtaining the logical sum with the brain image data of the white matter portion binarized in step 1 and the gray matter / white matter partial brain image data logical sum step The brain image data for which the logical sum is obtained is recorded in the embedding step for embedding a predetermined portion with one value, the brain image data subjected to the embedding operation in the embedding step, and the identification brain image data recording unit. All-white matter and all-white matter discriminating brain image data for identifying all gray matter and all white matter are created and recorded in the discriminating brain image data recording unit. A gray matter / whole white matter logical product step, wherein the whole gray matter discrimination brain image data creation step is configured to perform a predetermined processing on the brain image data of the white matter portion binarized in the white matter partial brain image data binarization step. A white matter portion embedding step of performing an operation of embedding a portion with one value, and the white matter portion from the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording portion; A whole gray matter difference step of creating a whole gray matter identification brain image data by recording the brain image data of the white matter portion subjected to the embedding operation in the embedding step and recording it in the discrimination brain image data recording unit. be able to.

According to the intracranial volume and local brain structure analysis program of the present invention, the brain image data of the subject's brain measured by the computer tomography method is first subjected to predetermined preprocessing, and at least the brain image data of the gray matter portion Is obtained and recorded in the brain image data recording unit after separation. Next, a region of interest is set in the brain image data of the standard brain, and the brain image data of the standard brain in which the region of interest is set is recorded in the region of interest recording unit. The brain image data of the standard brain recorded in the region of interest recording unit is converted into the brain image data of the gray matter portion in the standard brain and the brain image data of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit. Based on the transformation, the region of interest is converted into the brain image data of the subject's brain and recorded as brain image data of the subject's brain in the region of interest recording unit. As described above, according to the intracranial volume and local brain structure analysis program of the present invention, desired ROIs are set in advance on the brain image data of the standard brain, and the standard brain in which the ROIs are set are set. The brain image data can be transformed into brain image data of the subject's brain. As a result, a desired region of interest can be set on the brain image data of the subject's brain with high accuracy regardless of the operator's experience or the like. As a result, ROIs can be set automatically and accurately not only at the level of the whole brain, cerebrum, or cerebellum but also at a finer anatomical region of the brain and used practically in clinical settings. There is an effect that can be.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 shows an overview of data used as appropriate in each embodiment of the present invention. Details of each data will be described later, but for the convenience of explanation, first, the relation of data to be used will be described. In FIG. 1, reference numeral 1 denotes a first brain image data recording unit for recording brain image data 3 of a subject's brain measured by a computer tomography method in a first format, and 2 denotes brain image data 3 of the subject's brain. It is a second brain image data recording unit that records brain image data 3AN of the subject's brain as a result of conversion into a predetermined format (“AN” is the first two letters of “Analyze” described later). From the brain image data 3AN of the subject's brain, the brain image data 3GM of the gray matter (GM) portion, the brain image data 3WM of the white matter (WM) portion, the cerebrospinal fluid (CSF) portion The brain image data 3CSF is separated and taken out, and is recorded in the brain image data recording unit 5 after the separation. In each embodiment of the present invention, the region of the brain to be processed is recorded in the region of interest recording unit 6. As shown in FIG. 1, the region-of-interest recording unit 6 has an intracranial region of interest 6 ic (“ic” is an acronym for intra cranial) and a local region of interest 6 VOI. On the region of interest side 6ic, the brain image data 7ic of the standard brain in which the intracranial region of interest is set and the brain image data 3ic of the subject's brain that has been subjected to a deformation process described later on the brain image data 7ic of the standard brain are provided. It is recorded. Similarly, in the local region of interest 6VOI, the brain image data 7VOI of the standard brain in which the local region of interest is set and the brain image data of the subject's brain that has been subjected to deformation processing described later on the brain image data 7VOI of the standard brain 3 VOI is recorded. By using the brain image data 3ic of the subject's brain recorded in the region-of-interest recording unit 6 and the brain image data 3GM of the gray matter portion recorded in the brain image data recording unit 5 after separation, the inside of the skull is identified. Intracranial identifying brain image data 9ic, all gray matter and all white matter identifying brain image data 9ALL ("ALL" means "all") and all gray matter identifying all gray matter Identification brain image data 9GM is obtained and recorded in the identification brain image data recording unit 8. By using the identification brain image data recorded in the identification brain image data recording unit 8 and the brain image data recorded in the post-separation brain image data recording unit 5, a voxel value such as in the skull can be obtained, The volume of the desired region of interest can be determined from the voxel values. In the present specification, each recording unit may be a database or a simple folder. Each image data can be a file recorded in a desired format.

Hereinafter, the first embodiment will be described with reference to the drawings while referring to the recording units and the data shown in FIG. FIG. 2 is a flowchart showing the overall flow of the intracranial volume and local brain structure analysis program or method according to the first embodiment of the present invention. As shown in FIG. 2, first, the brain image data 3 of the subject's brain measured by the computer tomography is subjected to predetermined preprocessing, and brain image data of at least the gray matter portion is acquired to obtain the separated brain image. It records in the data recording part 5 (pre-processing step, step S10). Here, an MRI image that is an anatomical image is suitable as a brain image measured by a computed tomography method, but a brain image measured using a CT image and a radiation diagnostic method can also be used. For example, a single-photon emission computed tomography that uses a radioisotope (RI), which is a brain image used in nuclear medicine clinical settings, and then uses γ-rays emitted from it. A SPECT image obtained by imaging (single photon emission computed tomography: SPECT), a PET image obtained by positron emission tomography (PET), or the like may be used. The SPECT image and the PET image are functional images representing functions such as blood flow before and after injection of a drug or a test drug.

  Next, a region of interest is set in the brain image data of the standard brain, and the brain image data 7ic of the standard brain in which the region of interest is set is recorded in the region of interest recording unit 6 (region of interest setting step, step S12). The region of interest is planarly set to a slice of brain image data (region of interest: ROI), and a three-dimensional 3D ROI (volume of interest: VOI) is set by collecting ROIs of a plurality of slices. Specific examples will be described later. In the following, the region of interest is referred to as “ROIs” unless otherwise distinguished.

  The brain image data 7ic and / or 7VOI of the standard brain recorded in the region of interest recording unit 6 is recorded in the brain image data of the gray matter portion in the standard brain and the gray matter in the subject's brain recorded in the brain image data recording unit 5 after separation. By transforming based on the brain image data 3GM of the portion, ROIs are converted into the brain image data of the subject's brain and recorded as the brain image data 3ic and / or 3VOI of the subject's brain in the region of interest recording unit 6 ( Region of interest conversion step (step S14).

In the standard brain, the entire intracranial morphology and the intracerebral positional information of the local region such as the hippocampus are standardized for each slice, so it is also possible for the operator to set ROIs in the local region such as the hippocampus. This can be performed with high accuracy regardless of the above. Therefore, in the intracranial volume and local brain structure analysis program of the present invention, desired ROIs are set in advance on the brain image data of the standard brain, and the brain image data of the standard brain on which the ROIs are set is set on the subject's brain. Transforms into brain image data. As a result, it is possible to set desired ROIs on the brain image data of the subject's brain with high accuracy regardless of the operator's experience or the like. In normal anatomical normalization, the brain image data of the subject's brain is deformed (normalized) so as to match the brain image data of the standard brain, but in the intracranial volume and local brain structure analysis program etc. of the present invention, In addition, the brain image data of the standard brain in which ROIs are set is transformed into brain image data of the subject's brain. Since the SPECT image and the PET image, which are functional images, are images that are totally blurred as compared with the MRI image, which is an anatomical image, the positional information in the brain is poor. For this reason, experience is required to set ROIs on a SPECT image. By using the intracranial volume and local brain structure analysis program of the present invention, the desired ROIs are set in advance on the brain image data of the standard brain, and the brain image data of the standard brain in which the ROIs are set It can be transformed into a subject's SPECT brain image or PET brain image. As a result, it is possible to set desired ROIs on the SPECT brain image or PET brain image of the subject with high accuracy regardless of the operator's experience or the like.

Hereinafter, the function of each part in the intracranial volume and local brain structure analysis program of the present invention will be described. FIG. 3 is a flowchart showing detailed processing of the preprocessing step (step S10) in the first embodiment of the present invention. As shown in FIG. 3, first, brain image data 3 of the subject's brain, which is measured by the computer tomography method and recorded in the first brain image data recording unit 1 in the first format, is converted into the second format. Is recorded in the second brain image data recording unit 2 (format conversion step, step S20). Here, as the first format, for example, DICOM (Digital imaging and communication in) which is a unified standard for medical image network transfer established by the ACR-NEMA (American College of Radiology-National Electrical Manufactures Association) standardization committee. medicine) format. Of course, the format for each medical device manufacturer may be used.

  FIG. 4 shows the DICOM format format 20. As shown in FIG. 4, each record 1 or the like includes a 2-byte group number 21, a 2-byte element number 22, a 2-byte data type 23, a 2-byte data length 24, and a data length. 24 is composed of data 25 having a size defined by 24. As shown in the example of FIG. 4, when the set of the group number 21 and the element number 22 is (0x0028, 0x0010), the data type 23 indicates Rows, that is, the matrix size in the y-axis direction. When the group of the group number 21 and the element number 22 is (0x0028, 0x0011), the data type 23 indicates Columns, that is, the matrix size in the x-axis direction. When the set of the group number 21 and the element number 22 is (0x7FE0, 0x0010), the data type 23 indicates Pixel Data, that is, image data. The structure of the image data is the same as the analyze format (second format) described later.

  Next, the second format will be described. An Analyze format can be used as the second format. FIG. 5 shows a three-dimensional space 30 for explaining the analysis format. Assume that a target 31 exists in a three-dimensional space 30 composed of an x-axis, a y-axis, and a z-axis in the direction as shown in FIG.

  FIG. 6 illustrates a program 40 that defines an analyze format. Although the program 40 is described in the c language, it is needless to say that the program 40 may be described in another programming language. As shown in FIG. 6, the program 40 defines an image data file (* .img) 41 and a header file (* .hdr) 42, and the definition of each axis shown in FIG. The data type “data” is described by the program unit 44.

  FIG. 7 shows the structure of the image data file 41. As shown in FIG. 7, data 43x in the x-axis direction is recorded from left to right in FIG. These x-axis direction data 43x are connected as shown by arrows from top to bottom in FIG. 7 to form y-axis direction data 43y, that is, data for one slice. The data 43y for one slice is recorded from the top to the bottom in FIG. 7, and constitutes data 43z in the z-axis direction, that is, data for all slices.

  FIG. 8 shows the structure of the header file 42. As shown in FIG. 8, 2 bytes from the 0th byte of the data address of the header file 42 record the size 45 of the header file 42, and 6 bytes from the 42nd byte indicate the matrix size 46 of the image data on the x-axis. 2 bytes each are recorded in the order of the matrix size in the direction, the matrix size in the y-axis direction, and the matrix size in the z-axis direction. The 2 bytes from the 70th byte record the type 47 of the image data, and 6 bytes from the 78th byte. The byte records the pixel size 48 of the image data in the order of the pixel size in the x-axis direction, the pixel size in the y-axis direction, and the pixel size in the z-axis direction, and 2 bytes from the 347th byte are the end of the file ( EOF) 49 is recorded. As described above, the header file 42 is composed of 348 bytes. FIG. 9 shows a schematic display 51 of the brain image data 3AN of the subject's brain converted into the analyze format. In the schematic display 51, the upper left shows a coronal section image (coronal) of the brain image data 3AN of the subject's brain, the upper right shows a sagittal section image (sagittal), and the lower left shows a transverse section image (axial).

  Next, the head of the brain is identified for the brain image data 3AN of the subject's brain recorded in the second brain image data recording unit 2 in the format conversion step (step S20), and the subjects within a predetermined range from the identified top of the head. Slice cutting for extracting brain image data of the brain is performed (extraction step, step S22). In general, since data of about 18 cm or more from the top of the head is data of the neck, the predetermined range is preferably about 18 cm from the top of the head. FIG. 10 schematically shows processes 51 and 52 for slicing and cutting the brain image data 3AN of the subject's brain. Similarly to the schematic display 51, the schematic display 52 schematically showing the result of slicing and cutting shows a coronal cross-sectional image at the upper left, a sagittal cross-sectional image at the upper right, and a horizontal cross-sectional image at the lower left.

  In general, it is difficult to completely fix a subject's head for a long time during imaging, and the subject's head may tilt. For this reason, the anterior comissure-posterior comissure (AC-PC) line in the brain image data of the subject's brain extracted in the extraction step (step S22) is used as the AC-PC line in the brain image data of the standard brain. Registration (registration) is performed (image registration step, step S24).

  FIG. 11 schematically shows processing for registering the brain image data of the subject's brain extracted in the extraction step (step S22) using the standard brain of the MNI coordinate system as a reference by using the schematic displays 52, 53, and 54. Similarly to the schematic display 51 in the schematic display 53 showing the standard brain of the MNI (Montreal Neurological Institute) coordinate system and the schematic display 54 schematically showing the result of image superposition, the upper left is the coronal cross-sectional image. The upper right shows a sagittal section image, and the lower left shows a transverse section image.

  Image superposition (registration) is a conversion from one space to another. In the first embodiment, the brain image data of the subject's brain extracted in the extraction step (step S22) is converted into the anatomical normal of SPM (Statistical parametric mapping) 99 with reference to the standard brain (MNI atlas) of the MNI coordinate system. The AC-PC line of the subject's brain is matched with the AC-PC line of the standard brain by performing rigid body transformation only by linear transformation by conversion. The standard brain of the MNI coordinate system is the Full Width Half Maximum in the Digital Phantom that can be downloaded from Brain Web: Simulated Brain Database (http://www.bic.mni.mcgill.ca/brainweb/). : FWHM) Use 6 mm Gaussian smoothing. Gaussian smoothing is one type of smoothing that weights and averages the values distributed in the vicinity of a pixel by weighting the values distributed in the vicinity of the pixel so as to reduce the distant part.・ Use filters. SPM99 is free software, and the latest version of SPM2 may be used. Alternatively, 3D-SSP (three-dimensional stereotactic surface projections) can be used. The reason for performing rigid body transformation only by linear transformation (parallel movement, rotation) is to prevent the brain image data of the subject's brain from being deformed. Brain image data (schematic display 54) that has been processed to match the AC-PC line is recorded as a matrix size of 182 × 218 × 182 and a voxel size of 1 mm × 1 mm × 1 mm.

  FIG. 12 shows six parameters used in the rigid transformation using only linear transformation in the image superimposing step (step S24). In FIG. 12, the brain 60 of the subject is in a three-dimensional space indicated by the X axis, the Y axis, and the Z axis. In order to register the image of the brain 60 of the subject, in addition to a vector in three directions of a translation dx in the X direction, a translation dy in the Y direction, and a translation dz in the Z direction, rotation around the X axis Six parameters having three rotation components, dθx, rotation dθy around the Y axis, and rotation dθz around the Z axis, are used. The difference between the result of the rotation / translation using the six parameters and the standard brain of the MNI coordinate system subjected to the Gaussian smoothing shown by the schematic display 53 is taken, and the difference is evaluated. As a result of the evaluation, if it is determined that the predetermined convergence condition is satisfied, brain image data matching the AC-PC line as shown by the schematic display 54 can be obtained, and the registration ends. On the other hand, if it is not determined that the convergence condition is satisfied, the six parameters are updated and the rotation / translation is repeated. As the above evaluation and the predetermined convergence condition, it is possible to use an evaluation or the like such that the mean square sum of the differences is not more than a desired small value.

  The brain image data 3GM of at least the gray matter portion is separated from the brain image data (schematic representation 54) of the subject's brain subjected to the image superposition in the image superposition step (step S24), and the post-separation brain image data is recorded. Record in part 5 (separation step, step S26). The segmentation algorithm of SPM99 can separate the brain image data 3AN of the subject's brain into a gray matter portion, a white matter portion, and a cerebrospinal fluid portion. 13A shows gray matter brain image data 3GM separated by the SPM99 segmentation algorithm, and FIG. 13B shows white matter brain image data 3WM separated by the algorithm. The drawing about the brain image data 3CSF of the separated cerebrospinal fluid portion is omitted. The gray matter brain image data 3GM and the like are all images having a voxel value of 0.0 to 1.0. For example, in the brain image data 3GM of the gray matter portion shown in FIG. 13A, the voxel value of the voxel completely filled with the gray matter portion is 1.0. Similarly, in the brain image data 3WM of the white matter portion shown in FIG. 13B, the voxel value of the voxel completely filled with the white matter portion is 1.0. On the other hand, the voxel value of the voxel at the boundary between the gray matter portion and the white matter portion is determined according to the ratio of each voxel. For example, in a voxel having a gray matter portion of 56% and a white matter portion of 44%, the voxel value of the gray matter portion is 0.56 and the voxel value of the white matter portion is 0.44. In the separation step (step S26) in the first embodiment, it is possible to separate at least the brain image data 3GM of the gray matter portion by using the segmentation algorithm of SPM99. Other software such as 3D-SSP can also be used. The brain image data of the other part is used in Example 2 and the like to be described later, but the brain image data of the other part is not particularly used in Example 1.

The coordinates of ROIs set in the brain image data of the standard brain in the region of interest setting step (step S12) described above are defined in advance for each part in the standard brain and recorded in a definition file or the like. Can do. The intracranial volume and local brain structure analysis program according to the present invention automatically reads the coordinates corresponding to the region from the definition file or the like based on the name of the region designated by the operator, such as the hippocampus. Can be set. Of course, the operator can manually set ROIs in an arbitrary region with a mouse or the like. As a combination of both, the operator can designate a site by name, and the operator can manually set ROIs in an arbitrary region with a mouse or the like. Once set, ROIs in an arbitrary region can be recorded in a definition file or the like with an appropriate name. FIG. 14A shows an example in which local ROIs are set in the hippocampal region as a cross-sectional image 70. In FIG. 14 (A), from the top of the head downward, the upper left slice SL1 is shown in order to the upper right slice SL5, followed by the lower left slice SL6 to the lower right slice SL9, and the symbol Hr is the hippocampus of the right brain ROI and Hl set in the region are ROIs set in the hippocampus region of the left brain. As shown in FIG. 14A, the ROI (Hr) set for the right brain and the ROI (H1) set for the left brain are displayed from slices SL3 to SL8, and these ROIs are gathered to gather ROIs in the hippocampus region. (VOI) is set. FIG. 14B shows a cross-sectional image 71 of an example in which local ROIs are set in a brain stem region. In FIG. 14B, slices SL1 to SL9 are shown downward from the top of the head in the same order as in FIG. 14A, and symbol Bs is an ROI set in the brain stem. As shown in FIG. 14B, the ROI (Bs) set for the brainstem is displayed from slices SL2 to SL8, and these ROIs are gathered to set the ROIs (VOI) for the brainstem. FIG. 14C shows an example in which local ROIs are set in an arbitrary region by a cross-sectional image 72. In FIG. 14C, slices SL1 to SL8 are shown from the top to the bottom in the same order as in FIG. 14A. Symbol Ar is ROI set in an arbitrary region of the right brain, and Al is the left brain. ROI set in an arbitrary area. As shown in FIG. 14 (C), ROI (Ar) is displayed from slices SL2 to SL5, while ROI (Al) is displayed from slices SL4 to SL7. ROIs (VOI) in the area are set.

  As described above, the brain image data 3GM of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit 5 in the region-of-interest conversion step (step S14) uses the segmentation algorithm of SPM99 as described above. It is the separated brain image data. FIG. 15 shows the processing of the region of interest conversion step (step S14) in the first embodiment of the present invention. In FIG. 15, the parts denoted by the same reference numerals as those in FIG. In FIG. 15, reference numeral 80 denotes brain image data of the gray matter portion of the standard brain of the MNI coordinate system that can be downloaded by Brain Web. As shown in FIG. 15, the brain image data 7VOI and / or 7ic of the standard brain is converted into brain image data 80 of the gray matter portion of the standard brain of the MNI coordinate system and brain image data 3GM of the gray matter portion in the subject's brain. The ROIs are converted into the brain image data of the subject's brain, and are recorded as the brain image data 3VOI and / or 3ic of the subject's brain in the region-of-interest recording unit 6. The deformation can be performed using the SPM99 anatomical normalization algorithm. In this case, the brain image data 3GM of the gray matter portion in the subject's brain is used as a template. That is, the subject's brain is regarded as a standard brain, and the standard brain in which ROIs are set is deformed to match the subject's brain. The brain image data 3VOI of the subject's brain to which the local ROIs are converted and the brain image data 3ic of the subject's brain to which the intracranial ROIs are converted are 0 when the voxel value is 0, and 1 when the voxel value is greater than 0 2 It is set as the valued image data.

  Instead of using the brain image data 3GM of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit 5 in the region of interest conversion step (step S14), the gray matter portion subjected to predetermined smoothing is used. The brain image data 3GM ′ can be used. In this case, the brain image data 3GM ′ of the gray matter portion subjected to predetermined smoothing is recorded in the post-separation brain image data recording unit 5 instead of the brain image data 3GM. FIG. 16 shows a process in which the brain image data 3GM of the gray matter portion in the subject's brain is subjected to predetermined smoothing to obtain the brain image data 3GM ′ of the gray matter portion. As the predetermined smoothing, Gaussian smoothing with an FWHM of 8 mm can be used. FIG. 17 shows the processing of the region of interest conversion step (step S14) when the brain image data 3GM 'of the gray matter portion of the subject's brain subjected to predetermined smoothing is used. In FIG. 17, portions denoted by the same reference numerals as those in FIG. As shown in FIG. 17, the brain image data 7VOI and / or 7ic of the standard brain is converted into the brain image data 80 of the gray matter portion of the standard brain of the MNI coordinate system and the brain image data 3GM of the gray matter portion in the subject's brain. The ROIs are transformed into the brain image data of the subject's brain by transforming based on the brain image data 3GM ′ subjected to the predetermined smoothing, and the brain image data 3VOI of the subject's brain is stored in the region of interest recording unit 6 and / or Or record as 3 ic. The deformation can be performed using the SPM99 anatomical normalization algorithm. In this case, brain image data 3GM ′ obtained by performing predetermined smoothing on the brain image data 3GM of the gray matter portion in the subject's brain is used as a template.

As described above, according to the first embodiment of the present invention, the brain image data 3 of the subject's brain measured by the computer tomography is first subjected to predetermined preprocessing, and at least the brain image data of the gray matter portion is obtained. Obtained and recorded in the separated brain image data recording unit 5. Here, an MRI image that is an anatomical image is suitable as the brain image measured by the computed tomography, but a CT image, a SPECT image, a PET image, or the like can also be used. Next, ROIs are set in the brain image data of the standard brain, and the brain image data 7 ic of the standard brain in which the ROIs is set is recorded in the region of interest recording unit 6. The brain image data 7ic and / or 7VOI of the standard brain recorded in the region of interest recording unit 6 is recorded in the brain image data of the gray matter portion in the standard brain and the gray matter in the subject's brain recorded in the brain image data recording unit 5 after separation. By transforming based on the partial brain image data 3GM, the ROIs are converted into the brain image data of the subject's brain and recorded as the brain image data 3ic and / or 3VOI of the subject's brain in the region-of-interest recording unit 6. As described above, in the intracranial volume and local brain structure analysis program of the present invention, the desired ROIs are set in advance on the brain image data of the standard brain, and the brain image of the standard brain in which the ROIs are set. Transform the data into brain image data of the subject's brain. As a result, it is possible to set desired ROIs on the brain image data of the subject's brain with high accuracy regardless of the operator's experience or the like. That is, in the intracranial volume and local brain structure analysis program of the present invention, the brain image data of the standard brain set with ROIs is transformed into the brain image data of the subject's brain, contrary to normal anatomical normalization. is doing. As a result, ROIs can be set automatically and accurately not only at the level of the whole brain, cerebrum, or cerebellum but also at a finer anatomical region of the brain and used practically in clinical settings. It is possible to provide an intracranial volume and a program for analyzing a local brain structure .

  In the second embodiment, by using the feature of the MRI image as an anatomical image representing the morphology of the brain, it is recorded in the region of interest recording unit 6 in the region of interest conversion step (step S12) in the first embodiment. A volume calculating step of calculating the volume of the ROIs based on the ROIs converted into the brain image data 3ic of the subject's brain can be further provided. Therefore, the computed tomography in the second embodiment uses MRI.

  Along with the volume calculation step, the preprocessing step (step S10), the region of interest setting step (step S12), and the region of interest conversion step (step S14) in the first embodiment are changed as follows. The brain image data acquired in the pre-processing step (step S10) and recorded in the post-separation brain image data recording unit 5 are the gray matter portion brain image data 3GM, the white matter portion brain image data 3WM, and the cerebrospinal cord, respectively. It is assumed that the brain image data 3CSF of the liquid part is included. As described in the first embodiment, the SPM99 segmentation algorithm can separate the brain image data 3AN of the subject's brain into a gray matter portion, a white matter portion, and a cerebrospinal fluid portion. In the first embodiment, only the brain image data of the gray matter portion is used, but in the second embodiment, three are used together. In the region of interest setting step (step S12), the intracranial ROIs of the entire intracranial region are set in the brain image data of the standard brain, and the brain image data (7ic) of the standard brain is recorded in the region of interest recording unit 6. In the region of interest conversion step (step S14), the brain image data 7ic of the standard brain set with the intracranial ROIs recorded in the region of interest recording unit 6 is converted into the brain image data 80 of the gray matter portion in the standard brain of the MNI coordinate system. The ROIs are converted into the brain image data of the subject's brain by transforming based on the brain image data 3GM or 3GM ′ of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit 5 The area of interest recording unit 6 records the brain image data 3 ic of the subject's brain.

  In the second embodiment, calculation of the volume of intracranial ROIs will be described. First, the overall flow of the volume calculation step will be described, and then each process will be described in detail. FIG. 18 is a flowchart showing the overall flow of the volume calculation step in the second embodiment of the present invention. As shown in FIG. 18, first, by using the brain image data 3GM or the like recorded in the post-separation brain image data recording unit 5, the intracranial ROIs recorded in the region of interest recording unit 6 are converted. Intracranial identification brain image data 9ic for identifying intracranial ROIs is created from the brain brain image data 3ic and recorded in the identification brain image data recording unit 8 (intracranial identification brain image data creation step, step S40).

  Next, the gray matter brain image data 3GM and the white matter brain image data 3WM recorded in the post-separation brain image data recording unit 5 and the intracranial identification brain image data recorded in the identification brain image data recording unit 8 are recorded. Based on 9 ic, all gray matter and all white matter discrimination brain image data 9ALL for identifying all gray matter and all white matter is created and recorded in the discrimination brain image data recording unit 8 (all gray matter and all white matter discrimination brain image data Creating step, step S42).

  Based on the whole gray matter / whole white matter discrimination brain image data 9ALL recorded in the discrimination brain image data recording unit 8 and the brain image data 3WM of the white matter portion recorded in the separated brain image data recording unit 5, The whole gray matter identification brain image data 9GM to be identified is created and recorded in the discrimination brain image data recording unit 8 (total gray matter discrimination brain image data creation step, step S44).

  It is determined whether to calculate the volume of the intracranial ROIs (step S46). As described above, in Example 2, the process proceeds to step S48 in order to calculate the volume of the intracranial ROIs. Otherwise, the process proceeds to the process indicated by the out-of-page connector A (Example 3, see FIG. 25). In step S48, the volume of the intracranial ROIs is calculated by multiplying the total voxel value in the intracranial ROIs identified by the intracranial identification brain image data 9ic recorded in the identification brain image data recording unit 8 by a predetermined numerical value. As the predetermined numerical value, 1/1000 is preferable. As explained in the first embodiment, the brain image data 3GM of the gray matter portion is an image having a voxel value of 0.0 to 1.0, and the voxel of the voxel completely filled with the gray matter portion or the like. The value is 1.0. Therefore, the total voxel value of voxels completely filled with gray matter or the like matches whether the total number of voxels is a real number or an integer. However, since the total voxel value includes the voxel value of the voxel partially filled with the gray matter portion or the like, it does not coincide with the total number of voxels. However, if the counting method of the voxels partially filled (for example, voxel value = 0.6) is filled with the gray matter portion or the like is counted at the filled ratio (for example, 0.6), the total voxel value Corresponds to the total number of voxels. The total voxel value in this specification and the like may be paraphrased as the total number of voxels (real number) counted as described above.

  Whole gray matter / whole white matter discrimination brain image data 9ALL recorded in the discrimination brain image data recording unit 8, brain image data 3GM of the gray matter portion and brain image data of the white matter portion recorded in the separated brain image data recording unit 5 Based on 3WM, the total voxel value in the whole gray matter and the whole white matter region is obtained, and the total voxel value is multiplied by a predetermined value to calculate the volume of the whole gray matter and the whole white matter region (total gray matter Total white matter region volume calculating step (step S50). As the predetermined numerical value, 1/1000 is preferable.

  Brain image data 3GM of the gray matter portion recorded in the post-separation brain image data recording unit 5 in the whole gray matter region identified by the whole gray matter identification brain image data 9GM recorded in the identification brain image data recording unit 8 The total value of the voxel values in the corresponding voxels is obtained, and the total gray matter area volume is calculated by multiplying the total value by a predetermined numerical value (total gray matter area volume calculating step, step S52). As the predetermined numerical value, 1/1000 is preferable.

  The total gray matter / total white matter region volume calculating step (step 50) and the total gray matter region volume calculating step (step 52) from the total gray matter and the total voxel values in the total white matter region, By subtracting, the total voxel value in the entire white matter region is obtained, and the volume of the entire white matter region is calculated by multiplying the total voxel value by a predetermined numerical value (total white matter region volume calculating step, step S54). As the predetermined numerical value, 1/1000 is preferable.

  Next, details of each process will be described. FIG. 19 is a flowchart showing the detailed flow of the intracranial identification brain image data creation step (step S40), and FIG. 20 shows the process of the intracranial identification brain image data creation step (step S40) using brain image data. 19 and 20, first, the gray matter portion brain image data 3GM, the white matter portion brain image data 3WM, and the cerebrospinal fluid portion brain image data recorded in the post-separation brain image data recording unit 5 3CSF is added (brain image data addition step, step S60).

  Next, the brain image data added in the brain image data addition step (step S60) is subjected to predetermined smoothing (brain image data smoothing step, step S62). As the predetermined smoothing, Gaussian smoothing with a FWHM of 6 mm can be used.

  Subsequently, the brain image data smoothed in the brain image data smoothing step (step S62) is binarized based on a predetermined threshold (brain image data binarization step, step S64). As the predetermined threshold, for example, 40% of the maximum value of the voxel value in the smoothed brain image data can be used, and the voxel value in the brain image data is 40% or more of the maximum value. A binary image can be obtained by setting 1 in some cases and 0 otherwise. The threshold value of 40% of the maximum value is optional and can be changed depending on the image quality of the MRI image at each facility.

  Logical product of the brain image data binarized in the brain image data binarization step (step S64) and the brain image data 3ic of the subject's brain whose intracranial ROIs are converted in the region of interest conversion step (step S14) ( AND), intracranial identification brain image data 9ic for identifying intracranial ROIs is created and recorded in the identification brain image data recording unit 8 (intracranial logical product step, step S66). Although it is theoretically sufficient to use the brain image data 3ic as the intracranial identification brain image data for identifying the intracranial ROIs, in reality, there may be noise in the brain image data 3ic. Therefore, by taking the logical product of the brain image data 3ic and the brain image data binarized in the brain image data binarization step (step S64), the accurate intracranial identification brain image data 9ic from which noise is removed is obtained. Can be obtained.

  FIG. 21 is a flowchart showing the detailed flow of the whole gray matter / whole white matter discrimination brain image data creation step (step S42), and FIG. 22 shows the processing of the whole gray matter / whole white matter discrimination brain image data creation step (step S42). Shown by brain image data. As shown in FIGS. 21 and 22, first, binarization is performed on the gray matter brain image data 3GM recorded in the post-separation brain image data recording unit 5 based on a predetermined threshold (gray matter partial brain image). Data binarization step (step S70). As the predetermined threshold, for example, 65% of the maximum value of the voxel value in the brain image data 3GM of the gray matter portion can be used, and the voxel value in the brain image data 3GM of the gray matter portion has a maximum value of 65%. A binary image can be obtained by setting 1 if it is% or more, and 0 otherwise. The threshold value of 65% of the maximum value is optional and can be changed depending on the image quality of the MRI image at each facility.

  Similarly, the white matter brain image data 3WM recorded in the post-separation brain image data recording unit 5 is binarized based on a predetermined threshold (white matter partial brain image data binarization step, step S72). As the predetermined threshold, for example, 35% of the maximum value of the voxel value in the brain image data 3WM of the white matter portion can be used, and the voxel value in the brain image data 3WM of the white matter portion is 35% or more of the maximum value. A binary image can be obtained by setting 1 if it is and 0 otherwise. The threshold value of 35% of the maximum value is optional and can be changed depending on the image quality of the MRI image at each facility.

  Subsequently, the gray matter brain image data binarized in the gray matter partial brain image data binarization step (step S70) and the binarized in the white matter partial brain image data binarization step (step S72). A logical sum with the brain image data of the white matter portion is obtained (gray matter / white matter partial brain image data logical sum step, step S74).

  An operation of embedding a predetermined portion with one value is performed on the brain image data for which the logical sum is obtained in the gray matter / white matter partial brain image data logical sum step (step S74) (embedding step, step S76). The predetermined portion is a portion of a small gap that looks like a so-called worm-eaten state inside the brain image data for which the logical sum is obtained.

  By calculating the logical product of the brain image data subjected to the embedding operation in the embedding step (step S74) and the intracranial identification brain image data 9ic recorded in the identification brain image data recording unit 8, total gray matter and total white matter are obtained. Is generated and recorded in the discrimination brain image data recording unit 8 (all gray matter / all white matter logical product step, step S78). It is theoretically sufficient to use the brain image data subjected to the embedding operation (step S74) as the whole gray matter / white matter discrimination brain image data for identifying all gray matter and all white matter. . However, in practice, there may be noise in the brain image data. Therefore, by taking the logical product of the brain image data subjected to the embedding operation in the embedding step (step S74) and the intracranial identification brain image data 9ic, the entire gray matter / white matter identification brain image data 9ALL from which noise has been removed is obtained. Can be obtained.

  FIG. 23 is a flowchart showing the detailed flow of the whole gray matter identification brain image data creation step (step S44), and FIG. 24 shows the processing of the whole gray matter discrimination brain image data creation step (step S44) using brain image data. As shown in FIG. 23 and FIG. 24, first, for the white matter brain image data binarized in the white matter partial brain image data binarization step (step S72), an operation of embedding a predetermined portion with one value is performed. (Step for embedding white matter part. Step S80). The predetermined portion is a small gap portion in a so-called worm-eaten state inside the binarized white matter brain image data. These portions are considered to be white matter portions because they are zero values.

  Subsequently, the brain image data of the white matter portion subjected to the embedding operation in the white matter portion embedding step (step S80) is subtracted from the all gray matter / whole white matter discrimination brain image data 9ALL recorded in the discrimination brain image data recording unit 8. By doing so, the whole gray matter identification brain image data 9GM is created and recorded in the discrimination brain image data recording unit 8 (total gray matter difference step, step S82). As described above, the whole gray matter identification brain image data 9GM for identifying the whole gray matter portion can be created.

  Next, the details of the all gray matter / total white matter region volume calculating step (step S50) will be described. In the outline portion of the all gray matter / whole white matter discrimination brain image data 9ALL recorded in the discrimination brain image data recording section 8 (see FIG. 22, the white portion at the edge of the brain image), it is recorded in the post-separation brain image data recording section 5. The larger voxel value of the corresponding voxel values of the gray matter portion brain image data 3GM and the white matter portion brain image data 3WM is used, while the whole gray matter / whole white matter distinguishing brain image data 9ALL By using the voxel value 1 in the inside (see FIG. 22, gray portion inside the brain image), the total voxel value in the whole gray matter and the whole white matter region is obtained. The total voxel value is multiplied by a predetermined numerical value (preferably 1/1000 as described above) to calculate the total gray matter and the volume of the whole white matter region.

As described above, according to the second embodiment of the present invention, the region of interest conversion step (step S12) in the first embodiment is performed by using the feature of the MRI image as the anatomical image representing the brain form. The volume calculation step of calculating the volume of the intracranial ROIs based on the ROIs converted into the brain image data 3ic of the subject's brain recorded in the region-of-interest recording unit 6 can be further provided. In the volume calculation step, not only the intracranial volume but also the total gray matter volume and the total white matter volume can be obtained. In the second embodiment, when the intracranial identification brain image data 9 ic is obtained, the logical product of the brain image data 3 ic and the brain image data binarized in the brain image data binarization step (step S 64) is taken. Therefore, accurate intracranial identification brain image data 9ic from which noise is removed can be obtained. When obtaining the whole gray matter / whole white matter identification brain image data 9ALL and the like, the logical product of the brain image data subjected to the embedding operation in the embedding step (step S74) and the intracranial identification brain image data 9ic is taken. Further, it is possible to obtain all gray matter / whole substance discrimination brain image data 9ALL from which noise is removed. As described above, the volume of the set intracranial ROIs can be automatically and accurately measured, and the intracranial volume and local brain structure analysis program etc. that can be used practically in the clinical field are provided. can do.

  In the third embodiment, calculation of the volume of the local ROIs will be described. This is basically the same as in the second embodiment, and the region of interest is recorded in the region of interest conversion step (step S12) in the first embodiment by using the features of the MRI image as an anatomical image representing the brain form. Based on the ROIs converted into the brain image data 3VOI of the subject's brain recorded in the unit 6, the volume of the ROIs is calculated.

  With the calculation of the volume of the local ROIs, the region-of-interest setting step (step S12) and the region-of-interest conversion step (step S14) in the second embodiment are changed as follows. In the region-of-interest setting step (step S12), the entire intracranial ROIs are set in the brain image data of the standard brain, and some local ROIs in the skull are set in the brain image data of another standard brain. Then, the brain image data (7ic and 7VOI) of each standard brain is recorded in the region of interest recording unit 6. In the region of interest conversion step (step S14), the brain image data 7ic of the standard brain set with the intracranial ROIs recorded in the region of interest recording unit 6 is converted into the brain image data 80 of the gray matter portion in the standard brain of the MNI coordinate system. The ROIs are converted into the brain image data of the subject's brain by transforming based on the brain image data 3GM or 3GM ′ of the gray matter portion in the subject's brain recorded in the post-separation brain image data recording unit 5 The area of interest recording unit 6 records the brain image data 3 ic of the subject's brain. Furthermore, the brain image data 7VOI of the standard brain set with the local ROIs recorded in the region of interest recording unit 6 is transferred to the brain image data 80 of the gray matter portion in the standard brain of the MNI coordinate system and the brain image data recording unit 5 after separation. The ROIs are transformed into the brain image data of the subject's brain by transforming based on the recorded gray matter brain image data 3GM or 3GM 'in the subject's brain, and the subject's brain is stored in the region of interest recording unit 6. Brain image data of 3 VOI.

  When it is determined in step S46 of the process flowchart (see FIG. 18) of the second embodiment that the volume of the intracranial ROIs is not calculated, that is, when the volume of the local ROIs is calculated, the process proceeds to the process indicated by the out-of-page connector A. . FIG. 25 is a flowchart showing the flow of processing for calculating the volume of local ROIs according to the third embodiment of the present invention. As shown in the flowchart from the out-of-page connector A in FIG. 25, first, all the gray matter / white matter discrimination brain image data creation step (step S42) described above and recorded in the discrimination brain image data recording unit 8 are recorded. The gray matter / whole white matter identification brain image data 9ALL and the area of the whole gray matter and the whole white matter, and the local ROIs recorded in the region of interest recording unit 6 in the region of interest conversion step (step S14) are converted. Based on the gray matter brain image data 3GM and the white matter brain image data 3WM recorded in the post-separation brain image data recording unit 5 in the common area with the local ROIs indicated by the brain brain image data 3VOI, Find the total voxel value of gray matter and white matter in ROIs. Next, the gray matter and white matter volumes in the local ROIs are calculated by multiplying the total voxel value by a predetermined numerical value (local gray matter / white matter volume calculating step, step S90). As the predetermined numerical value, 1/1000 is preferable. As shown in FIG. 14 (A) or FIG. 14 (C), when there is a left-right distinction in the local ROIs, values are calculated separately on the left and right.

  Subsequently, the region of the whole gray matter identified by the whole gray matter discrimination brain image data 9GM created in the all gray matter discrimination brain image data creation step (step S44) and recorded in the discrimination brain image data recording unit 8, and the interest The post-separation brain image data recording unit 5 in the common region with the local ROIs indicated by the brain image data 3VOI of the subject's brain converted from the local ROIs recorded in the region of interest recording unit 6 in the region conversion step (step S14). The total value of the voxel values in the corresponding voxels of the gray matter portion brain image data 3GM recorded in (1) is obtained. The gray matter volume in the local ROIs is calculated by multiplying the total value by a predetermined numerical value (local gray matter volume calculating step, step S92). As the predetermined numerical value, 1/1000 is preferable. As shown in FIG. 14 (A) or FIG. 14 (C), when there is a left-right distinction in the local ROIs, values are calculated separately on the left and right.

  By subtracting the total value obtained in the local gray matter volume calculating step (step 92) from the total voxel value of the gray matter and white matter in the local ROIs obtained in the local gray matter / white matter volume calculating step (step S90). The total voxel value in the white matter region of the local ROIs is obtained. The volume of white matter in the local ROIs is calculated by multiplying the total voxel value by a predetermined value (local white matter region volume calculating step, step S94). As the predetermined numerical value, 1/1000 is preferable. As shown in FIG. 14 (A) or FIG. 14 (C), when there is a left-right distinction in the local ROIs, values are calculated separately on the left and right.

  Next, details of the above-mentioned local gray matter / white matter volume calculating step (step S90) will be described. First, the whole gray matter and the whole white matter region identified by the whole gray matter / whole white matter discrimination brain image data 9ALL recorded in the discrimination brain image data recording unit 8 and the local ROIs recorded in the region of interest recording unit 6 are In the common area with the local ROIs indicated by the converted brain image data 3VOI of the subject's brain, the outline portion of the common area (see FIG. 22, the white portion at the edge of the brain image) is stored in the post-separation brain image data recording unit 5. Among the voxel values of the voxels corresponding to the recorded gray matter portion brain image data 3GM and white matter portion brain image data 3WM, the larger voxel value is used, while the inside of the common region (see FIG. 22). In the gray portion of the image), the voxel value 1 is used to obtain the gray matter in the local ROIs and the total voxels in the white matter region. By multiplying the total voxel value by a predetermined numerical value (preferably 1/1000), the gray matter and white matter volume in the local ROIs can be calculated. As shown in FIG. 14 (A) or FIG. 14 (C), when there is a left-right distinction in the local ROIs, values are calculated separately on the left and right.

As described above, according to the third embodiment of the present invention, basically, as in the second embodiment, the feature of the MRI image as an anatomical image representing the morphology of the brain is used. Based on the ROIs converted into the brain image data 3VOI of the subject's brain recorded in the region-of-interest recording unit 6 in the region-of-interest conversion step (step S12) in Example 1, the volume of the local ROIs can be calculated. Not only the volume of the entire local ROIs but also the total gray matter volume and the total white matter volume of the local ROIs can be determined. As described above, the volume of local ROIs set in an arbitrary region can be automatically and accurately measured, and the intracranial volume and local brain structure analysis program can be practically used in the clinical field. Etc. can be provided.

Next, experiments conducted to verify the intracranial volume and local brain structure analysis program of the present invention will be described.

Methods and subjects.
The subjects were 15 Alzheimer's disease patients and 15 normal elderly persons, and three-dimensional brain MRI was performed. From the obtained data, the intracranial volume, total brain volume, and hippocampal volume were manually determined, and the intracranial volume and local brain structure of the present invention were used. Measurement was performed with a physical analysis program (hereinafter referred to as “the present method”), and a correlation coefficient between the two was obtained.

result.
The correlation coefficient r between this method and manual measurement was as follows: intracranial volume r = 0.910, whole brain volume r = 0.902, hippocampal volume r = 0.918.

Conclusion.
This method can measure the intracranial volume, total brain volume, and hippocampal volume in a short time and fully automatically, and is considered useful for clinical use.

Specific target.
15 Alzheimer's disease patients (average age 63.4 ± 6.9 years, male: female = 4: 11) and 15 age-matched normal volunteers (average age 61.3 ± 5.9 years, male: female = 4:11).

Specific method.
As MRI imaging, head MRI was performed using a 1.5 Tesla MRI apparatus. Axis T2-weighted image and coronal 3D-SPGR image (TR 14 ms, TE 3 ms, flip angle 20 degrees, matrix size 256 × 256, pixel size 0.86 × 0.86 mm, slice thickness 1. 5 mm × 124 slices) were imaged.

analysis.
The coronal section 3D-SPGR image was converted into an analyze format and then reconstructed into an axial section with a pixel size of 1.0 mm × 1.0 mm and a slice thickness of 1.0 mm. The intracranial volume was manually excised by identifying the boundary with the dura mater and cerebrospinal fluid for each slice. For the total brain volume, the volume obtained by excluding the cerebrospinal cavity with the signal value of average of cerebrospinal fluid × 0.8 signal value as a threshold on the cut image was defined as the total brain volume. Hippocampal volume is determined by the method of Jack et al. (Jack CR Jr, Twomey CK, Zinsmeister AR, Sharbrough FW, Petersen RC, Cascino GD: Anterior temporal lobes and hippocampal formations: Normative volumetric measurements from MR images in young adults. Radiology 1989; 172: 549 -554.) Using manual and segmentation methods.

  Correlation coefficients of the intracranial volume, total brain volume, and hippocampal volume of 30 persons obtained by this method and manually were obtained to verify the quantitativeness. In addition, the presence or absence of significant differences between the intracranial volume, the whole brain volume, and the hippocampal volume of the Alzheimer's disease group and the normal group obtained by this method and the manual group was verified using a t-test.

  I will supplement the method of Jack et al. The hippocampus boundary is traced to include Ammon horn, hippocampal white board, dentate gyrus, and hippocampus. The number of pixels in which the pixel value is between the maximum value of cerebral gray matter and the maximum value of cerebrospinal fluid to the maximum value of cerebral gray matter is counted, and the volume of the hippocampus is multiplied by the pixel size and slice thickness. Calculated.

result.
It took about 4 hours per case to manually calculate the intracranial volume, total brain volume, and hippocampal volume. In this method, measurement was possible in about 10 minutes.

  The correlation coefficient r between this method and manual measurement was intracranial volume r = 0.910, whole brain volume r = 0.902, hippocampal volume r = 0.918. FIG. 26 shows the intracranial volumes of the Alzheimer's disease group and normal group obtained by each method. In FIG. 26, TIV is the intracranial volume, Whole brain is the whole brain volume, Hippocampi is the hippocampal volume, AD is an Alzheimer's disease patient, NC is a normal volunteer, Manual is a manual measurement, and AMSIS is a measurement by this method. The unit of the numerical value is cc or ml. 27 to 29 show the correlation between this method and manual measurement for TIV (intracranial volume), Hippocampi (hippocampal volume), and whole brain (whole brain volume), respectively. 27 to 29, the vertical axis represents this method (AMSIS), the horizontal axis represents manual measurement (manual), and the unit is cc or ml for both axes.

Discussion.
The anatomical standardization method of SPM is a method in which individual brains of different shapes and sizes are combined with the standard brain. Since the volume is measured by fitting a ROI template set in advance on the standard brain to each person's brain, this method can be said to be a change of mind using the anatomical standardization method of SPM.

  The fully automatic volume measurement method like this method has the characteristics that there is no variation by the measurer as in the case of manual operation, and the same result is always obtained regardless of the experience of the measurer. In manual measurement, it took about 4 hours per case to measure the intracranial volume, the whole brain volume, and the hippocampal volume, but it can be done in only 10 minutes. In addition, since this method can measure small structures such as the hippocampus with the same accuracy as the intracranial volume and the whole brain volume, it can be applied to actual clinical diagnosis.

  This method uses the anatomical standardization method and segmentation method of SPM99, and the accuracy of the anatomical standardization and segmentation depends on SPM, and it is expected that the accuracy will be further improved by improving SPM in the future. . For segmentation only, the method of Lemieux et al. (Lemieux L, Hammers A, Mackinnon T, Liu RS. Automatic segmentation of the brain and intracranial cerebrospinal fluid in T1-weighted volume MRI scans of the head, and its application to serial cerebral and Intracranial volumetry. Magn Reson Med. 2003; 49: 872-84.) is also possible.

  In this experimental example, the total brain volume and hippocampal volume were obtained, but as long as the brain template was prepared in advance, the gray matter, white matter, cerebrospinal cavity / ventricle, and other structures could be separated and measured. Is also possible.

  In the early stage of Alzheimer's disease, atrophy of the medial temporal lobe including the hippocampus is seen, but by measuring this volume it is possible to objectively know how much it is atrophy, and the volume of the normal person An early diagnosis can be made by comparison. Since this method can measure the volume in about 10 minutes, it can be implemented as a routine of daily medical care.

  If this method is installed in advance in the operation computer of each company's MRI apparatus, the volume of the target region can be calculated 10 minutes after the MRI imaging is completed.

  In addition, this method can locally measure the blood flow and metabolism of SPECT / PET by superimposing MRI and SPECT / PET images. Have.

FIG. 30 is a block diagram showing an internal circuit 90 of a computer that executes the computer program ( intracranial volume and local brain structure analysis program) of the present invention. As shown in FIG. 30, the CPU 91, ROM 92, RAM 93, image control unit 96, controller 98, input control unit 100 and external interface (Interface: I / F) unit 101 are connected to a bus 102. In FIG. 30, the above-described computer program of the present invention is recorded on a recording medium (including a removable recording medium) such as a ROM 92, a disk 97a or a CD-ROM 97n. The recording medium such as the disk 97a or the CD-ROM 97n includes the first brain image data recording unit 1, the second brain image data recording unit 2, the post-separation brain image data recording unit 5, the region of interest recording unit 6, and the discriminating brain. An image data recording unit 8 and the like are recorded.

  The computer program of the present invention is loaded into the RAM 93 from the ROM 92 via the bus 102 or from a recording medium such as the disk 97a or CD-ROM 97n via the controller 98 via the bus 102. The image control unit 96 sends brain image data to the VRAM 95, and the display unit 94 is a display device such as a display that displays a brain image based on the brain image data sent from the VRAM 95. The VRAM 95 is an image memory having a capacity corresponding to the data capacity of one screen of the display unit 94. The input operation unit 99 is an input device such as a mouse or a numeric keypad for inputting to the computer. The input control unit 100 is connected to the input operation unit 99 and performs input control and the like. The external I / F unit 101 has an interface function when connecting to an external communication network (not shown) such as the Internet.

  As described above, the CPU 91 executes the computer program of the present invention to achieve the object of the present invention. The computer program can be supplied to the computer CPU 91 in the form of a recording medium such as a CD-ROM 97n as described above, and a recording medium such as a CD-ROM 97n on which the computer program is recorded also constitutes the present invention. It will be. As the recording medium on which the computer program is recorded, for example, a memory card, a memory stick, a DVD, an optical disk, an FD, or the like can be used in addition to the recording medium described above.

As an application example of the present invention, it can be applied to an actual clinical diagnosis for a small brain structure such as the hippocampus. Since volume measurement is possible in a short time (about 10 minutes), it can be applied to routines of daily medical care. It is possible to locally measure the blood flow / metabolic amount of SPECT / PET by superimposing the MRI in which ROIs are set by the intracranial volume and local brain structure analysis program of the present invention and the SPECT or PET image. Since it becomes possible, it can be applied to various fields.

It is a figure which shows the whole image of the data used suitably in each Example of this invention. It is a flowchart which shows the whole flow of the intracranial volume and local brain structure analysis program or method in Example 1 of this invention. It is a flowchart which shows the detailed process of the pre-processing step (step S10) in Example 1 of this invention. It is a figure which shows the DICOM format format 20. FIG. It is a figure which shows the three-dimensional space 30 for demonstrating an analyze format. It is a figure which illustrates the program 40 which defines an analysis format. 3 is a diagram showing a configuration of an image data file 41. FIG. 3 is a diagram illustrating a configuration of a header file 42. FIG. It is a figure which shows the model display of the brain image data 3AN of the test subject's brain converted into the analyze format. It is a figure which shows the process which slices and cuts with respect to the brain image data 3AN of a test subject's brain with typical display 51 and 52. FIG. It is a figure which shows the process which performs image superimposition on the brain image data of a test subject's brain extracted by the extraction step (step S22) on the basis of the standard brain of a MNI coordinate system by typical display 52, 53, and 54. It is a figure shown about six parameters used by rigid transformation only of linear transformation in a picture superposition step (Step S24). It is a figure which shows the brain image data 3GM of the gray matter part isolate | separated by the segmentation algorithm of SPM99. It is a figure which shows the brain image data 3WM of the white matter part isolate | separated by the segmentation algorithm of SPM99. It is a figure which shows by the cross-sectional image 70 the example which set local ROIs to the site | part of the hippocampus. It is a figure which shows the example which set local ROIs in the site | part of the brain stem with the cross-sectional image 71. FIG. It is a figure which shows the example which set local ROIs to arbitrary area | regions by the cross-sectional image 72. FIG. It is a figure which shows the process of the region of interest conversion step (step S14) in Example 1 of this invention. It is a figure which shows the process which gives predetermined smoothing to the brain image data 3GM of the gray matter part in a test subject's brain, and makes it the brain image data 3GM 'of a gray matter part. It is a figure which shows the process of a region of interest conversion step (step S14) at the time of using the brain image data 3GM 'of the gray matter part of the test subject's brain which performed predetermined smoothing. It is a flowchart which shows the flow of the volume calculation step in Example 2 of this invention. It is a flowchart which shows the detailed flow of an intracranial identification brain image data creation step (step S40). It is a figure which shows the process of an intracranial identification brain image data creation step (step S40) with brain image data. It is a flowchart which shows the detailed flow of an all gray matter and all white matter identification brain image data creation step (step S42). It is a figure which shows the process of a whole gray matter and all white matter identification brain image data creation step (step S42) by brain image data. It is a flowchart which shows the detailed flow of a whole gray matter identification brain image data creation step (step S44). It is a figure which shows the process of a whole gray matter identification brain image data creation step (step S44) by brain image data. It is a flowchart which shows the flow of the process which calculates the volume of the local ROIs in Example 3 of this invention. It is a figure which shows the intracranial volume of the Alzheimer's disease group and the normal group obtained by this method and manual measurement. It is a figure which shows the correlation with this method and manual measurement about TIV (intracranial volume). It is a figure which shows the correlation with this method and manual measurement about Hippocampi (hippocampal volume). It is a figure which shows the correlation with this method and manual measurement about Whole brain (whole brain). It is a block diagram which shows the internal circuit 210 of the computer which performs the computer program of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 1st brain image data recording part, 2 2nd brain image data recording part, 3, 3AN, 3ic, 3VOI Brain image data of a test subject's brain, 3GM, 3GM 'Gray matter part brain image data, 3WM White matter part brain Image data, 3CSF brain image data of cerebrospinal fluid portion, 5 post-separation brain image data recording unit, 6, 6ic, 6VOI region of interest recording unit, 7ic, 7VOI brain image data of standard brain, 8 identification brain image data recording unit, 9ic intracranial identification brain image data, 9ALL all gray matter / white matter identification brain image data, 9GM all gray matter identification brain image data, 20 DICOM format, 21 group number, 22 element number, 23 data type, 24 data length, 25 Data, 30 3D space, 31 object, 40 program, 41 image data file, 42 header File, 43 definition part, 44 program part, 43x x-axis direction data, 43y y-axis direction data, 43z z-axis direction data, 45 header file 42 size, 46 image data matrix size, 47 image data type , 48 pixel size of image data, 49 end of file (EOF), 51, 52, 53, 54 schematic representation of brain image data 3AN of subject's brain, 60 subject's brain, 70, 71, 72 cross-sectional image, Brain image data of gray matter portion of standard brain of 80 MNI coordinate system, 90 internal circuit, 91 CPU, 92 ROM, 93 RAM, 95 VRAM, 96 image control unit, 97a disk, 97n CD-ROM, 98 controller, 99 inputs Operation unit, 100 input control unit, 101 external I / F unit, 102 bus.

Claims (21)

An intracranial volume and local brain structure analysis program for causing a computer to execute processing of brain image data, comprising:
An extraction step of extracting brain image data of the subject's brain in a predetermined range for the brain image data of the subject's brain measured by computer tomography and recorded in the brain image data recording unit;
An image superposition step for performing image superposition for matching the anterior commissure-posterior commissural line in the brain image data of the subject's brain extracted in the extraction step to the vicinity of the anterior commissure-posterior commissural line in the brain image data of the standard brain;
When the intracranial volume is not calculated from the brain image data of the subject's brain that has undergone image superposition in the image superimposing step, at least the gray matter brain image data is separated and the post-separated brain image data recorded Separation step to record in the part,
A region-of-interest setting step in which a region of interest is set in the brain image data of the standard brain and recorded in the region-of-interest recording unit;
The brain image data of the standard brain recorded in the region of interest recording unit is the brain image data of the gray matter part in the standard brain and the brain image of the gray matter part in the subject's brain recorded in the post-separation brain image data recording unit. The brain image data of the standard brain in which the region of interest is set is subjected to the reverse anatomical normalization to match the brain image data of the subject, and into the brain image data of the subject's brain . An intracranial volume and local brain structure analysis program for executing a region of interest conversion step of converting the region of interest and recording it in the region of interest recording unit. In the intracranial volume and local brain structure analysis program according to claim 1 , prior to the extraction step,
A format in which brain image data of a subject's brain measured by a computer tomography method and recorded in a first brain image data recording unit in a first format is converted into a second format and recorded in a second brain image data recording unit A conversion step ,
The extraction step identifies the head of the brain image data of the subject's brain recorded in the second brain image data recording unit in the format conversion step, and the brain of the subject within a predetermined range from the identified top of the head. intracranial volume and local brain structure analysis program and extracting the brain image data. The intracranial volume and local brain structure analysis program according to claim 1 or 2, wherein the brain image of the gray matter portion in the brain of the subject recorded in the post-separation brain image data recording unit prior to the region of interest conversion step. An intracranial volume and local brain structure analysis program, further comprising a smoothing step of performing predetermined smoothing on data and recording the data in the post-separation brain image data recording unit. In the intracranial volume and local brain structure analysis program according to any one of claims 1 to 3,
The computed tomography in the preprocessing step is magnetic resonance imaging,
A volume calculation step of calculating a volume of the region of interest based on the region of interest converted in the brain image data of the subject's brain recorded in the region of interest recording unit in the region of interest conversion step; Characteristic intracranial volume and local brain structure analysis program. In the intracranial volume and local brain structure analysis program according to claim 4,
The brain image data acquired in the pre-processing step and recorded in the post-separation brain image data recording unit are the brain image data of the gray matter portion, the brain image data of the white matter portion, and the brain image of the cerebrospinal fluid portion, respectively. Data and
The region-of-interest setting step sets an intracranial region of interest in the entire skull within the brain image data of the standard brain, records the brain image data of the standard brain in the region of interest recording unit,
In the region of interest conversion step, the brain image data of the standard brain in which the intracranial region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the post-separated brain image data recording unit The intracranial region of interest is converted into the brain image data of the subject's brain and is recorded in the region of interest recording unit by being deformed based on the brain image data of the gray matter portion in the subject's brain recorded in And
The volume calculating step includes
By using the brain image data recorded in the post-separation brain image data recording unit, the intracranial region of interest is obtained from the brain image data of the subject's brain converted from the intracranial region of interest recorded in the region of interest recording unit. Intracranial identification brain image data creating step of creating intracranial identification brain image data to be identified and recording the identification brain image data recording unit;
Based on the gray matter brain image data and white matter brain image data recorded in the post-separation brain image data recording unit, and the intracranial identification brain image data recorded in the identification brain image data recording unit, A gray matter / whole white matter identification brain image data creating step for creating a whole gray matter / whole white matter identification brain image data for identifying gray matter and whole white matter and recording it in the identification brain image data recording unit;
Based on the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording section and the brain image data of the white matter portion recorded in the post-separation brain image data recording section, the total gray matter is identified. Creating a whole gray matter identification brain image data and recording it in the identification brain image data recording unit;
When calculating the volume of the region of interest within the skull,
Multiplying the total voxel value in the intracranial region of interest identified by the intracranial identification brain image data recorded in the discriminating brain image data recording unit by calculating a volume of the intracranial region of interest;
Whole gray matter / whole white matter identification brain image data recorded in the identification brain image data recording unit, gray matter portion brain image data and white matter portion brain image data recorded in the separated brain image data recording unit Based on the total gray matter and the total white matter region, a total voxel value in the whole gray matter and the total white matter region is obtained, and the total gray matter and the total white matter region are calculated by multiplying the total voxel value by a predetermined numerical value. A volume calculating step;
Corresponding to the brain image data of the gray matter portion recorded in the post-separation brain image data recording unit in the whole gray matter region identified by the whole gray matter identification brain image data recorded in the identification brain image data recording unit A total gray matter region volume calculating step of calculating a total value of voxel values in the voxels to be calculated and calculating a volume of the total gray matter region by multiplying the total value by a predetermined numerical value;
By subtracting the total value determined in the total gray matter region volume calculation step from the total voxel value in the total gray matter and total white matter region determined in the total gray matter / total white matter region volume calculation step, total white matter obtains the total voxel value in the region, intracranial volume, characterized in that a total white matter area volume calculation step of calculating the volume of the region of該全white matter by multiplying a predetermined value to said total voxel value and Local brain structure analysis program. The intracranial volume and local brain structure analysis program according to claim 5, wherein the total gray matter / total white matter region volume calculating step includes:
In the outline portion of the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, the brain image data of the gray matter portion and the brain image of the white matter portion recorded in the post-separation brain image data recording unit Using the larger voxel value of the voxel values corresponding to the data, and using the voxel value 1 in the whole gray matter / whole white matter identification brain image data, the whole gray matter and the whole white matter region A program for analyzing intracranial volume and a local brain structure , wherein the total voxel value is calculated and the total gray matter and the volume of the whole white matter region are calculated by multiplying the total voxel value by a predetermined numerical value. In the intracranial volume and local brain structure analysis program according to claim 5 or 6,
In the region-of-interest setting step, an entire intracranial region of interest within the skull is set in the brain image data of the standard brain, and a local region of interest within the skull is set in another brain image data of the standard brain. Record the brain image data of each standard brain in the region of interest recording unit,
In the region of interest conversion step, the brain image data of the standard brain in which the intracranial region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the post-separated brain image data recording unit The intracranial region of interest is converted into the brain image data of the subject's brain and is recorded in the region of interest recording unit by being deformed based on the brain image data of the gray matter portion in the subject's brain recorded in In addition, the brain image data of the standard brain in which the local region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the post-separated brain image data recording unit By transforming based on the brain image data of the gray matter portion in the subject's brain, the local region of interest is converted into the brain image data of the subject's brain and recorded in the region of interest recording unit,
The volume calculating step calculates the volume of the local region of interest,
Whole gray matter and whole white matter regions identified by the whole gray matter / whole white matter discrimination brain image data created in the whole gray matter / whole white matter discrimination brain image data creation step and recorded in the discrimination brain image data recording unit; In the common region with the local region of interest indicated by the brain image data of the subject's brain converted from the local region of interest recorded in the region of interest recording unit in the region of interest conversion step, the post-separation brain image data recording unit Based on the recorded gray matter portion brain image data and white matter portion brain image data, a total voxel value of gray matter and white matter in the local region of interest is obtained, and the total voxel value is multiplied by a predetermined numerical value. A local gray matter / white matter volume calculating step for calculating gray matter and white matter volume in the local region of interest;
A region of total gray matter identified by the whole gray matter identification brain image data created in the whole gray matter identification brain image data creation step and recorded in the discrimination brain image data recording unit, and the region of interest in the region of interest conversion step The brain image of the gray matter portion recorded in the post-separation brain image data recording unit in the common region with the local region of interest indicated by the brain image data of the subject's brain converted from the local region of interest recorded in the recording unit Determining a total value of voxel values in the voxel corresponding to the data, and calculating a gray matter volume in the local region of interest by multiplying the total value by a predetermined numerical value; and
By subtracting the total value determined in the local gray matter volume calculation step from the total gray matter and white matter voxel values in the local interest region determined in the local gray matter / white matter volume calculation step, obtains the total voxel value in the area of white matter, characterized by further comprising a local white matter region volume calculation step of calculating the white matter volume in said local region of interest by multiplying the predetermined number to said total voxel value cranium Internal volume and local brain structure analysis program. The intracranial volume and local brain structure analysis program according to claim 7, wherein the local gray matter / white matter volume calculating step includes:
The whole gray matter and the whole white matter region identified by the whole gray matter / whole white matter discrimination brain image data recorded in the discriminating brain image data recording unit and the local region of interest recorded in the region of interest recording unit are converted. In the common area with the local region of interest indicated by the brain image data of the subject's brain, the brain image data of the gray matter portion and the brain of the white matter portion recorded in the post-separation brain image data recording portion in the outline portion of the common region The voxel value of the voxel corresponding to the image data is used, and the voxel value 1 is used inside the common region, thereby obtaining the total voxels in the gray matter and white matter regions in the local region of interest. determined, intracranial volume and topical and calculates the gray matter and white matter of the volume in said local region of interest by multiplying the predetermined number in the aggregate voxel value Structural analysis program. The intracranial volume and local brain structure analysis program according to any one of claims 5 to 8, wherein the intracranial identification brain image data creation step includes:
A brain image data addition step of adding the brain image data of the gray matter portion, the brain image data of the white matter portion and the brain image data of the cerebrospinal fluid portion recorded in the post-separation brain image data recording unit;
A brain image data smoothing step for applying a predetermined smoothing to the brain image data added in the brain image data adding step;
A brain image data binarization step for binarizing the brain image data smoothed in the brain image data smoothing step based on a predetermined threshold;
By calculating the logical product of the brain image data binarized in the brain image data binarizing step and the brain image data of the subject's brain whose intracranial region of interest has been converted in the region of interest conversion step, An intracranial volume and local brain structure analysis program, comprising: an intracranial logical product step of creating intracranial identification brain image data for identifying a region of interest and recording it in the discriminating brain image data recording unit. In the intracranial volume and local brain structure analysis program according to any one of claims 5 to 9,
The whole gray matter / whole white matter identification brain image data creation step includes:
A gray matter partial brain image data binarization step for binarizing the gray matter brain image data recorded in the post-separation brain image data recording unit based on a predetermined threshold;
A white matter partial brain image data binarization step for binarizing the brain image data of the white matter portion recorded in the post-separation brain image data recording unit based on a predetermined threshold;
Logic between the gray matter brain image data binarized in the gray matter partial brain image data binarization step and the white matter brain image data binarized in the white matter partial brain image data binarization step Gray matter / white matter partial brain image data logical sum step for sum,
An embedding step of performing an operation of embedding a predetermined portion with one value for the brain image data obtained by the logical sum in the gray matter / white matter partial brain image data logical step,
By taking the logical product of the brain image data subjected to the embedding operation in the embedding step and the intracranial identification brain image data recorded in the identification brain image data recording unit, all gray matter and all white matter are identified. A gray matter / whole white matter discrimination brain image data is created and recorded in the discrimination brain image data recording unit, and a gray matter / whole white matter AND step,
The whole gray matter identification brain image data creation step includes:
A white matter partial embedding step of performing an operation of embedding a predetermined portion with one value for the white matter brain image data binarized in the white matter partial brain image data binarization step;
By subtracting the white matter portion brain image data subjected to the embedding operation in the white matter portion embedding step from the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, An intracranial volume and local brain structure analysis program, comprising: a whole gray matter difference step for creating discrimination brain image data and recording the discrimination brain image data in the discrimination brain image data recording unit. The computer-readable recording medium which recorded the intracranial volume and local brain structure analysis program in any one of Claims 1 thru | or 10. An intracranial volume and local brain structure analysis method for causing a computer to process brain image data,
Extracting the brain image data of the subject's brain in a predetermined range for the brain image of the subject's brain measured by the computer tomography method and recorded in the brain image data recording unit;
An image superposition step for performing image superposition for matching the anterior commissure-posterior commissure line in the brain image data of the subject's brain extracted in the extraction step to the vicinity of the anterior commissure-posterior commissure line in the brain image data of the standard brain;
When the intracranial volume is not calculated from the brain image data of the subject's brain that has undergone image superposition in the image superimposing step, at least the gray matter brain image data is separated and the post-separated brain image data recorded A separation step to record in a part;
A region of interest setting step of setting a region of interest in the brain image data of the standard brain and recording it in the region of interest recording unit,
The brain image data of the standard brain recorded in the region of interest recording unit is the brain image data of the gray matter part in the standard brain and the brain image of the gray matter part in the subject's brain recorded in the post-separation brain image data recording unit. The brain image data of the standard brain in which the region of interest is set is subjected to the reverse anatomical normalization to match the brain image data of the subject, and into the brain image data of the subject's brain . A region-of-cranial volume and local brain structure analysis method comprising: a region-of-interest conversion step of converting the region of interest and recording the region of interest in a region-of-interest recording unit. The intracranial volume and local brain structure analysis method according to claim 12, wherein , prior to the extraction step,
A format in which brain image data of a subject's brain measured by a computer tomography method and recorded in a first brain image data recording unit in a first format is converted into a second format and recorded in a second brain image data recording unit A conversion step ,
The extraction step identifies the head of the brain image data of the subject's brain recorded in the second brain image data recording unit in the format conversion step, and the brain of the subject within a predetermined range from the identified top of the head. intracranial volume and local brain structure analysis method and extracting the brain image data. 14. The intracranial volume and local brain structure analysis method according to claim 12 or 13, wherein the brain image of the gray matter portion in the brain of the subject recorded in the post-separation brain image data recording unit prior to the region of interest conversion step. An intracranial volume and local brain structure analysis method, further comprising a smoothing step of performing predetermined smoothing on data and recording the data in the post-separation brain image data recording unit. The intracranial volume and local brain structure analysis method according to any one of claims 12 to 14,
The computed tomography in the preprocessing step is magnetic resonance imaging,
A volume calculation step of calculating a volume of the region of interest based on the region of interest converted in the brain image data of the subject's brain recorded in the region of interest recording unit in the region of interest conversion step; Characteristic method for analyzing intracranial volume and local brain structure . The intracranial volume and local brain structure analysis method according to claim 15,
The brain image data acquired in the pre-processing step and recorded in the post-separation brain image data recording unit are the brain image data of the gray matter portion, the brain image data of the white matter portion, and the brain image of the cerebrospinal fluid portion, respectively. Data and
The region-of-interest setting step sets an intracranial region of interest in the entire skull within the brain image data of the standard brain, records the brain image data of the standard brain in the region of interest recording unit,
In the region of interest conversion step, the brain image data of the standard brain in which the intracranial region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the post-separated brain image data recording unit The intracranial region of interest is converted into the brain image data of the subject's brain and is recorded in the region of interest recording unit by being deformed based on the brain image data of the gray matter portion in the subject's brain recorded in And
The volume calculating step includes
By using the brain image data recorded in the post-separation brain image data recording unit, the intracranial region of interest is obtained from the brain image data of the subject's brain converted from the intracranial region of interest recorded in the region of interest recording unit. Intracranial identification brain image data creating step of creating intracranial identification brain image data to be identified and recording the identification brain image data recording unit;
Based on the gray matter brain image data and white matter brain image data recorded in the post-separation brain image data recording unit, and the intracranial identification brain image data recorded in the identification brain image data recording unit, A gray matter / whole white matter identification brain image data creating step for creating a whole gray matter / whole white matter identification brain image data for identifying gray matter and whole white matter and recording it in the identification brain image data recording unit;
Based on the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording section and the brain image data of the white matter portion recorded in the post-separation brain image data recording section, the total gray matter is identified. Creating a whole gray matter identification brain image data and recording it in the identification brain image data recording unit;
When calculating the volume of the region of interest within the skull,
Multiplying the total voxel value in the intracranial region of interest identified by the intracranial identification brain image data recorded in the discriminating brain image data recording unit by calculating a volume of the intracranial region of interest;
Whole gray matter / whole white matter identification brain image data recorded in the identification brain image data recording unit, gray matter portion brain image data and white matter portion brain image data recorded in the separated brain image data recording unit Based on the total gray matter and the total white matter region, a total voxel value in the whole gray matter and the total white matter region is obtained, and the total gray matter and the total white matter region are calculated by multiplying the total voxel value by a predetermined numerical value. A volume calculating step;
Corresponding to the brain image data of the gray matter portion recorded in the post-separation brain image data recording unit in the whole gray matter region identified by the whole gray matter identification brain image data recorded in the identification brain image data recording unit A total gray matter region volume calculating step of calculating a total value of voxel values in the voxels to be calculated and calculating a volume of the total gray matter region by multiplying the total value by a predetermined numerical value;
By subtracting the total value determined in the total gray matter region volume calculation step from the total voxel value in the total gray matter and total white matter region determined in the total gray matter / total white matter region volume calculation step, total white matter obtains the total voxel value in the region, intracranial volume, characterized in that a total white matter area volume calculation step of calculating the volume of the region of該全white matter by multiplying a predetermined value to said total voxel value and Local brain structure analysis method. The intracranial volume and local brain structure analysis method according to claim 16, wherein the total gray matter / total white matter region volume calculating step includes:
In the outline portion of the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, the brain image data of the gray matter portion and the brain image of the white matter portion recorded in the post-separation brain image data recording unit Using the larger voxel value of the voxel values corresponding to the data, and using the voxel value 1 in the whole gray matter / whole white matter identification brain image data, the whole gray matter and the whole white matter region A method for analyzing intracranial volume and a local brain structure, comprising: calculating total voxel values and multiplying the total voxel value by a predetermined numerical value to calculate the volume of the total gray matter and the total white matter region. The intracranial volume and local brain structure analysis method according to claim 16 or 17,
In the region-of-interest setting step, an entire intracranial region of interest within the skull is set in the brain image data of the standard brain, and a local region of interest within the skull is set in another brain image data of the standard brain. Record the brain image data of each standard brain in the region of interest recording unit,
In the region of interest conversion step, the brain image data of the standard brain in which the intracranial region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the post-separated brain image data recording unit The intracranial region of interest is converted into the brain image data of the subject's brain and is recorded in the region of interest recording unit by being deformed based on the brain image data of the gray matter portion in the subject's brain recorded in In addition, the brain image data of the standard brain in which the local region of interest recorded in the region of interest recording unit is set, the brain image data of the gray matter portion in the standard brain and the post-separated brain image data recording unit By transforming based on the brain image data of the gray matter portion in the subject's brain, the local region of interest is converted into the brain image data of the subject's brain and recorded in the region of interest recording unit,
The volume calculating step calculates the volume of the local region of interest,
Whole gray matter and whole white matter regions identified by the whole gray matter / whole white matter discrimination brain image data created in the whole gray matter / whole white matter discrimination brain image data creation step and recorded in the discrimination brain image data recording unit; In the common region with the local region of interest indicated by the brain image data of the subject's brain converted from the local region of interest recorded in the region of interest recording unit in the region of interest conversion step, the post-separation brain image data recording unit Based on the recorded gray matter portion brain image data and white matter portion brain image data, a total voxel value of gray matter and white matter in the local region of interest is obtained, and the total voxel value is multiplied by a predetermined numerical value. A local gray matter / white matter volume calculating step for calculating gray matter and white matter volume in the local region of interest;
A region of total gray matter identified by the whole gray matter identification brain image data created in the whole gray matter identification brain image data creation step and recorded in the discrimination brain image data recording unit, and the region of interest in the region of interest conversion step The brain image of the gray matter portion recorded in the post-separation brain image data recording unit in the common region with the local region of interest indicated by the brain image data of the subject's brain converted from the local region of interest recorded in the recording unit Determining a total value of voxel values in the voxel corresponding to the data, and calculating a gray matter volume in the local region of interest by multiplying the total value by a predetermined numerical value; and
By subtracting the total value determined in the local gray matter volume calculation step from the total gray matter and white matter voxel values in the local interest region determined in the local gray matter / white matter volume calculation step, obtains the total voxel value in the area of white matter, characterized by further comprising a local white matter region volume calculation step of calculating the white matter volume in said local region of interest by multiplying the predetermined number to said total voxel value cranium Internal volume and local brain structure analysis method. The intracranial volume and local brain structure analysis method according to claim 18, wherein the local gray matter / white matter volume calculating step includes:
The whole gray matter and the whole white matter region identified by the whole gray matter / whole white matter discrimination brain image data recorded in the discriminating brain image data recording unit and the local region of interest recorded in the region of interest recording unit are converted. In the common area with the local region of interest indicated by the brain image data of the subject's brain, the brain image data of the gray matter portion and the brain of the white matter portion recorded in the post-separation brain image data recording portion in the outline portion of the common region By using the larger voxel value of the voxel values of the corresponding voxel with the image data and using the voxel value 1 inside the common region, the total voxel value in the gray matter and white matter regions in the local region of interest look, intracranial volume and station, characterized by calculating the gray matter and white matter of the volume in said local region of interest by multiplying the predetermined number in the aggregate voxel value Brain structure analysis method. The intracranial volume and local brain structure analysis method according to any one of claims 16 to 19, wherein the intracranial identification brain image data creation step includes:
A brain image data addition step of adding the brain image data of the gray matter portion, the brain image data of the white matter portion and the brain image data of the cerebrospinal fluid portion recorded in the post-separation brain image data recording unit;
A brain image data smoothing step for applying a predetermined smoothing to the brain image data added in the brain image data adding step;
A brain image data binarization step for binarizing the brain image data smoothed in the brain image data smoothing step based on a predetermined threshold;
By calculating the logical product of the brain image data binarized in the brain image data binarizing step and the brain image data of the subject's brain whose intracranial region of interest has been converted in the region of interest conversion step, A method for analyzing intracranial volume and local brain structure, comprising: an intracranial logical product step of creating intracranial identification brain image data for identifying a region of interest and recording it in the identification brain image data recording unit. The intracranial volume and local brain structure analysis method according to any one of claims 16 to 20,
The whole gray matter / whole white matter identification brain image data creation step includes:
A gray matter partial brain image data binarization step for binarizing the gray matter brain image data recorded in the post-separation brain image data recording unit based on a predetermined threshold;
A white matter partial brain image data binarization step for binarizing the brain image data of the white matter portion recorded in the post-separation brain image data recording unit based on a predetermined threshold;
Logic between the gray matter brain image data binarized in the gray matter partial brain image data binarization step and the white matter brain image data binarized in the white matter partial brain image data binarization step Gray matter / white matter partial brain image data logical sum step for sum,
An embedding step of performing an operation of embedding a predetermined portion with one value for the brain image data obtained by the logical sum in the gray matter / white matter partial brain image data logical step,
By taking the logical product of the brain image data subjected to the embedding operation in the embedding step and the intracranial identification brain image data recorded in the identification brain image data recording unit, all gray matter and all white matter are identified. A gray matter / whole white matter discrimination brain image data is created and recorded in the discrimination brain image data recording unit, and a gray matter / whole white matter AND step,
The whole gray matter identification brain image data creation step includes:
A white matter partial embedding step of performing an operation of embedding a predetermined portion with one value for the white matter brain image data binarized in the white matter partial brain image data binarization step;
By subtracting the white matter portion brain image data subjected to the embedding operation in the white matter portion embedding step from the whole gray matter / whole white matter discrimination brain image data recorded in the discrimination brain image data recording unit, A method for analyzing intracranial volume and local brain structure, comprising: a whole gray matter difference step of creating discrimination brain image data and recording the discrimination brain image data in the discrimination brain image data recording unit.